SteganoAmor, Fake IP Scanner, MITRE Corporation Breach - April 14-20, 2024 - This Week in Security
Dharminder from the F5 SIRT team covers Security news of the week, SteganoAmor, Google Ads to distribute fake IP scanner, MITRE Corporation Breach and Vulnerability Disclosure Program for Defense Industrial BaseInSpectre, Rust/PANOS CVEs, X URL blunder and More-April 8-14, 2024-F5 SIRT-This Week in Security
Editor'sIntroduction Hello, Arvin is your editor for This Week in Security. As usual, I collected some interesting security news. Credit to the original articles. Intel processors are affected by a Native Branch History Injection (Native BHI) attack and the tool InSpectre, a tool that can find gadgets (code snippets that can serve as a jumping point to bypass sw and hw protections) in an OS kernel on vulnerable hardware. Spectre style attacks that abuses speculative execution on processors has been around for a while now. Intel updated their previous published article on "Branch History Injection and Intra-mode Branch Target Injection" guidance and included an "Additional Hardening Options" section. The silver lining in this, is the CVEs CVSS score are Medium severity. See the section snippets from the research paper of the researchers from VU Amsterdam that illustrates the use InSpectre tool. Rust has a critical CVE - CVE-2024-24576. It affects the Rust standard library, which was found to be improperly escaping arguments when invoking batch files on Windows using the library's Command API – specifically, std::process::Command. It is specific to the Windows OS cmd exe as it has complex parsing rules and allowed untrusted inputs to be safely passed to spawned processes. Next is a PAN OS Critical CVE, where it affects devices with firewall configurations with GlobalProtect gateway and device telemetry enabled. CVE-2024-3400 affects PAN-OS 10.2, PAN-OS 11.0 and PAN-OS 11, Updates to fully fix this CVE were made available from April 14. Refer tohttps://security.paloaltonetworks.com/CVE-2024-3400 Change Healthcare's worries on effects of a previous breach due to ALPHV ransomware group appears to be not over. Per the report, the victim organization was potentially "exit" scammed by ALPHV and is being pursued by the "contactor/affiliate" of the ransomware attack, RansomHub, demanding another round of ransom to be paid, else, they sell the exfiltrated data to the highest bidder. X/Twitter had an URL blunder where it converts anything with the string twitter in their site's tweets and then converts it to the letter X - example, netflitwitter[.]com will be converted to netflix[.]com. This behavior was reversed and back to usual, but X twitter[.]com URLs now properly converts to X[.]com. Lastly, a round up of issues from MS, Fortinet, SAP, Cisco, Adobe, Google/Android. As in previous TWIS editions, some of these news were a recurrence/follow up. In general, keep your systems up to date on software versions, secure access to them and allow only trusted users and applications to run. Implement layers of protections - updated AV/ED/XDR on Server and End User systems, Firewall/network segmentation rules/IPS to prevent further spread/lateral movement in the event of a ransomware attack (BIG-IP AFM have network firewall, IPS features that you can consider), a WAF to protect your web applications and APIs - BIG-IP ASM/Adv WAF, F5 Distributed Cloud Services, NGINX App Protect have security policy configuration and attack signatures that can mitigate known command injection techniques and other web exploitation techniques. End user security training and awareness, incident response and reporting will help an organization should that first phishing email reaches a target end user mailbox. If it feels "off" and looks suspicious, stop and ponder before clicking. I hope this edition of TWIS is educational. You can also read past TWIS editions and othercontent from the F5 SIRT , so check those out as well. Till next time! Rust rustles up fix for 10/10 critical command injection bug on Windows in std lib Programmers are being urged to update their Rust versions after the security experts working on the language addressed a critical vulnerability that could lead to malicious command injections on Windows machines. The vulnerability, which carries a perfect 10-out-of-10 CVSS severity score, is tracked as CVE-2024-24576. It affects the Rust standard library, which was found to be improperly escaping arguments when invoking batch files on Windows using the library's Command API – specifically, std::process::Command. "An attacker able to control the arguments passed to the spawned process could execute arbitrary shell commands by bypassing the escaping," said Pietro Albini of the Rust Security Response Working Group, who wrotethe advisory. The main issue seems to stem from Windows' CMD.exe program, which has more complex parsing rules, and Windows can't execute batch files without it, according to the researcher at Tokyo-based Flatt Security whoreported the issue. Albini said Windows' Command Prompt has its own argument-splitting logic that works differently from the usual Command::arg and Command::args APIs provided by the standard library, which typically allow untrusted inputs to be safely passed to spawned processes. "On Windows, the implementation of this is more complex than other platforms, because the Windows API only provides a single string containing all the arguments to the spawned process, and it's up to the spawned process to split them," said Albini. "Most programs use the standard C run-time argv, which in practice results in a mostly consistent way arguments are split. "Unfortunately it was reported that our escaping logic was not thorough enough, and it was possible to pass malicious arguments that would result in arbitrary shell execution." https://www.theregister.com/2024/04/10/rust_critical_vulnerability_windows/ It's 2024 and Intel silicon is still haunted by data-spilling Spectre Intel CPU cores remain vulnerable to Spectre data-leaking attacks, say academics at VU Amsterdam. We're told mitigations put in place at the software and silicon level by the x86 giant to thwart Spectre-style exploitation of its processors' speculative execution can be bypassed, allowing malware or rogue users on a vulnerable machine to steal sensitive information – such as passwords and keys – out of kernel memory and other areas of RAM that should be off limits. The boffins say they have developed a tool called InSpectre Gadget that can find snippets of code, known as gadgets, within an operating system kernel that on vulnerable hardware can be abused to obtain secret data, even on chips that have Spectre protections baked in. InSpectre Gadget was used, as an example, to find a way to side-step FineIBT, a security feature built into Intel microprocessors intended to limitSpectre-stylespeculative execution exploitation, and successfully pull off a Native Branch History Injection (Native BHI) attack to steal data from protected kernel memory. "We show that our tool can not only uncover new (unconventionally) exploitable gadgets in the Linux kernel, but that those gadgets are sufficient to bypass all deployed Intel mitigations," the VU Amsterdam teamsaidthis week. "As a demonstration, we present the first native Spectre-v2 exploit against the Linux kernel on last-generation Intel CPUs, based on the recent BHI variant and able to leak arbitrary kernel memory at 3.5 kB/sec." https://www.theregister.com/2024/04/10/intel_cpus_native_spectre_attacks/ fromhttps://download.vusec.net/papers/inspectre_sec24.pdf 2.2 Spectre v2 In 2018, the disclosure of Spectre [29] famously demonstratedhow speculation can be used to leak data across security domains. One variant presented in the paper, originally known asSpectre v2 or Branch Target Injection (BTI), shows how speculation of indirect branches can be used to transiently divertthe control flow of a program and redirect it to an attackerchosen location. The attack works by poisoning one of theCPU predictors, the Branch Target Buffer (BTB), which isused to decide where to jump on indirect branch speculation. Initially, mitigations were proposed at the software leveland, later, in-silicon mitigations such as Intel eIBRS [5] anARM CSV2 [12] were added to newer generations of CPUsto isolate predictions across privilege levels. 2.3 Branch History Injection In 2022, Branch History Injection (BHI) [13] showed that,despite mitigations, cross-privilege Spectre v2 is still possibleon latest Intel CPUs by poisoning the Branch History Buffer(BHB). Figure 1 provides a high-level overview of the attack. In summary, by executing a sequence of conditionalbranches (HA and HV ) right before performing a system call,an unprivileged attacker can cause the CPU to transientlyjump to a chosen target (TA) when speculating over an indirect call in the kernel (CV ). This happens because the CPUpicks the speculative target forCV from a shared structure, theBTB, that is indexed using both the address of the instructionand the history of previous conditional branches, which isstored in the Branch History Buffer (BHB). Finding the rightcombination of histories that will result in a collision can bedone with brute-forcing.To ensure the injected target, TA, contains a disclosure gadget, the original BHI attack relied on the presence of theextended Berkeley Packet Filter (eBPF), through which anunprivileged user can craft code that lives in the kernel. Figure 2: InSpectre gadget workflow. The analyst provides akernel image and a list of target addresses to InSpectre Gadget⃝1 , which performs in-depth inspection to find gadgets thatcan leak secrets and output their characteristics. The gadgetscan be filtered ⃝2 based on the available attacker-controlledregisters and the mitigations enabled, and used to craft Spectrev2 exploits against the kernel ⃝3 . Zero-day exploited right now in Palo Alto Networks' GlobalProtect gateways Palo Alto Networks on Friday issued a critical alert for an under-attack vulnerability in the PAN-OS software used in its firewall-slash-VPN products. The command-injection flaw, with an unwelcome top CVSS severity score of 10 out of 10, may let an unauthenticated attacker execute remote code with root privileges on an affected gateway, which to put it mildly is not ideal. It can, essentially, be exploited to take complete control of equipment and drill into victims' networks. Updates to fully fix this severe hole are due to arrive by Sunday, April 14, we're told. CVE-2024-3400affects PAN-OS 10.2, PAN-OS 11.0 and PAN-OS 11.1 firewall configurations with a GlobalProtect gateway and device telemetry enabled. Cloud firewalls, Panorama appliances, and Prisma Access are not affected, Palo Altosays. Zero-day exploitation of this vulnerability was detected on Wednesday by cybersecurity shop Volexity, on a firewall it was monitoring for a client. After an investigation determined that the firewall had been compromised, the firm saw another customer get hit by the same intruder on Thursday. "The threat actor, which Volexity tracks under the alias UTA0218, was able to remotely exploit the firewall device, create a reverse shell, and download further tools onto the device," the networks security management firm said in ablog post. "The attacker focused on exporting configuration data from the devices, and then leveraging it as an entry point to move laterally within the victim organizations." The intrusion, which begins as an attempt to install a custom Python backdoor on the firewall, appears to date back at least to March 26, 2024. Palo Alto Networks refers to the exploitation of this vulnerability as Operation MidnightEclipse, which at least is more evocative than the alphanumeric jumble UTA0218. The firewall maker says while the vulnerability is being actively exploited, only a single individual appears to be doing so at this point. mitigations include applying a GlobalProtect-specificvulnerability protection, if you're subscribed to Palo Alto's Threat Prevention service, or "temporarily disabling device telemetry until the device is upgraded to a fixed PAN-OS version. Once upgraded, device telemetry should be re-enabled on the device." It urged customers to follow the above security advisory and thanked the Volexity researchers for alerting the company and sharing its findings. ® https://www.theregister.com/2024/04/12/palo_alto_pan_flaw/ https://www.volexity.com/blog/2024/04/12/zero-day-exploitation-of-unauthenticated-remote-code-execution-vulnerability-in-globalprotect-cve-2024-3400/ https://unit42.paloaltonetworks.com/cve-2024-3400/ Change Healthcare faces second ransomware dilemma weeks after ALPHV attack Change Healthcare is allegedly being extorted by a second ransomware gang, mere weeks after recovering from an ALPHV attack. RansomHub claimed responsibility for attacking Change Healthcare in the last few hours, saying it had 4 TB of the company's data containing personally identifiable information (PII) belonging to active US military personnel and other patients, medical records, payment information, and more. The miscreants are demanding a ransom payment from the healthcare IT business within 12 days or its data will be sold to the highest bidder. "Change Healthcare and United Health you have one chance in protecting your clients data," RansomHub said. "The data has not been leaked anywhere and any decent threat intelligence would confirm that the data has not been shared nor posted. The org is alleged to have paid a $22 million ransom to ALPHV following the incident – a claim made by researchers monitoring a known ALPHV crypto wallet and one backed up by RansomHub. However, Change Healthcare has never officially confirmed this to be the case. If all of the claims are true, it means the embattled healthcare firm is deciding whether to pay a second ransom fee to keep its data safe. the prevailing theory among infosec watchers is that ALPHV pulled what's known as an exit scam after Change allegedly paid its ransom. While the ratios vary slightly between gangs, generally speaking, ransomware payments are split 80/20 – 80 percent for the affiliate that actually carried out the attack and 20 percent for the gang itself. It's believed that ALPHV took 100 percent of the alleged payment from Change Healthcare, leaving the affiliate responsible for the attack without a commission. Angry and searching for what they believed they were "owed," the affiliate is thought to have retained much of the data it stole and now switched allegiances to RansomHub in one last throw of the dice to earn themselves a payday, or so the theory goes. UnitedHealth, parent company of Change Healthcare,discloseda cybersecurity incident on February 22, saying at the time it didn't expect it to materially impact its financial condition or the results of its operations. It originally suspected nation state attackers to be behind the incident, but the ALPHV ransomware gang later claimed responsibility. Many of its systems were taken down as a result while it assessed and worked to remediate the damage. Hospitals and pharmacies reported severe disruption to services following the attack, with many unable to process prescriptions, payments, and medical claims. Cashflow issues also plagued many institutions, prompting the US government tointervene. The IT biz's data protection standards are soon to be subject to aninvestigationby the US healthcare industry's data watchdog, which cited the "unprecedented magnitude of this cyberattack" in its letter to Change. https://www.theregister.com/2024/04/08/change_healthcare_ransomware/ X fixes URL blunder that could enable convincing social media phishing campaigns Elon Musk's X has apparently fixed an embarrassing issue implemented earlier in the week that royally bungled URLs on the social media platform formerly known as Twitter. Users started noticing on Monday that X's programmers implemented a rule on its iOS app that auto-changedTwitter.comlinks that appeared in Xeets toX.com links. Attackers could feasibly copy legitimate web pages to steal credentials, or skip the trouble and simply use it as a malware-dropping tool, or any number of other possibilities. The potential for abuse here would be rife, given the number of legitimate, well-known brands most people would blindly trust. Netflix, Plex, Roblox, Clorox, Xerox – you get the picture. According to tests at Reg towers on Wednesday morning, the issue appears to have been reversed. Netflitwitter[.]com now reads as such, but Twitter.com is auto-changed to X.com.
Maintainers, Slowloris/2, Kobold Letters - April 1st - 7th, 2024 - F5 SIRT - This Week in Security
Introduction Hello again, Kyle Fox here. This week we have some shorter bits about things, in which I promise two more future articles, which I think means I am up to three non-TWIS articles in the pipeline. We have to talk about project maintainers again. We have all seen that one XKCD comic about dependency maintainers. The xz situation has resurfaced a common plea from Open Source maintainers: We need funds and help. I don't have any real deep commentary here, just a plea that companies heavily dependent on Open Source projects should consider giving back to the community by retaining internal SMEs who can help projects resolve issues by submitting bug fixes, contribute to those projects financially, and possibly consider hiring internal people to work on the major features they want out of these projects. Platforms like GitHub may be able to help by moderating discussions to keep project maintainers from being abused by users. And the community should work better at being a positive force for change. And the same goes for conferences, some of us spend lots of time working on all the little details so you can go to DEF CON, have parties to go to, things to hack and places to hack them in. Its easy to look at something like DEF CON and think that its just another industry conference and everyone is being paid to be there, but very few people are paid to be there. I will further discuss this soon in a post about the current DEF CON situation and venues. Is the HTTP/2 CONTINUATION Attack Just Slowloris/2? On April 3rd the industry got wind of a new attack on HTTP/2, this time you could consume resources by sending a steady stream of CONTINUATION frames, leaving the connection open and consuming resources. This came on the tail end of the HTTP/2 Rapid Reset attack, which consumed resources in an orthogonal way. If this attack sounds familiar, its because it is almost the same attack for HTTP/2 as the Slowloris attack was for HTTP/1.1. You could also compare it to the Slow POST attack as well. How Slowloris worked, for those who may have forgotten since 2009, is the attacker will send a HTTP/1.1 request to a webserver and then slowly send one header at a time, holding the connection open for a very long time with limited traffic. On susceptible webservers they would only need to send headers fast enough to keep the TCP connection from timing out, since the webserver does not have a timeout for the header stage of the request. The Slow POST attack is similar, but slowly sending chunks of POST data rather than headers, relying on the webserver not timing out on those. BIG-IP mitigated Slowloris by its normal behavior of buffering all the headers before forwarding a request to the backend servers. A limit on the number and/or size of headers allows further refinement of this mitigation. When mitigated, these attacks only generate at most an open connection on the backend with no request. This same behavior mitigated the HTTP/2 Rapid Reset attack and now mitigates the HTTP/2 CONTINUATION attack. As we can see from this, old attacks can become new ones when a new or significantly revised protocol comes along. This is why when working on new features F5 performs Threat Modelling Assessments to categorize possible new variations of old attacks or completely new attacks that may apply to a new feature, protocol or service and build in protections against those attacks. Display: none Strikes Again, Now in Email. A recent post over at Lutra Security called Kobold Letters has resurfaced an old trick with CSS, but this time in email. The basic TL;DR of this trick is using display: none attached to CSS in an email to hide text in the email until its forwarded or replied to. Email clients often will convert an email to plain text or try to convert the HTML and CSS slightly. This results in the ability to put blocks of text in divs or other selectable blocks that can be styled in CSS to hide them or otherwise change their display and appearance when they are forwarded or replied to. I don't know if this really changes much in the spear-phishing risk area, at this point organizations should have considerable controls in places to make sure that fund transfers are only acted on with clear verified approval and that the destinations of fund transfers are vetted and verified, not copied from some email and sent without checking. Fortunately in this case the vendors have been informed and they are working to provide solutions to this attack, so it may not be viable for very long. Are Bluetooth Discovery Attacks Drying Up? I don't have much to write here since I have not yet dove into the data that much, but the Bluetooth Discovery attacks that I talked about in December appear to not be as popular as they once were. I used Wall-of-Flippers at a few conventions in March to collect Flipper and Bluetooth Discovery Spam data, but it appears that not a whole lot of spamming was happening. Apple and Google Android have been working on mitigating these attacks, Apple having released several iOS updates to patch it. The lack of impact these days may be driving this trend. I do intend on bringing the Wall-of-Flippers to more events, and will be doing a bigger writeup on the device, the software and the data collected here on DevCentral in the coming month or two. Roundup Not a channel this time, but a single video by TwinkleTwinkie: Understanding & Making PCB Art. Google to delete records made from users using Incognito Mode in lawsuit settlement. Microsoft has announced how much it will cost to keep Windows 10 past the date they want you to move to Windows 11. No word on a better Windows 11 UI. Fake AI lawfirms are sending DMCA takedowns to generate SEA gains. (Original report) A recommendation from my recent trip to Las Vegas: Roberto's Taco Shop. Wi-Fi only works when its raining. This is a lesson in sometimes the observations, while absurd, are correct. Roku wants to insert ads in HDMI inputs? DEF CON now has hotel blocks at the Sahara Las Vegas, The Fontainebleau Las Vegas and Resort World.CPU vulns, LoopDOS, AI worms and Bad Bots - March 18th - 24th - This Week in Security
Editor's introduction Aaron back with you as editor this week, and I'm going to look at a few of the interesting pieces of secrurity news from the week just gone - the good news is, there was a lot of interesting news! The bad news is, that interesting news usually means more work for us all... allow me to walk you through a few of the interesting things I saw this week and give you my take on what their real-world impact is going to be to us all. We in F5 SIRT invest a lot of time understanding the frequently changing behavior of bad actors. Bad actors are a threat to your business, your reputation, and your livelihood. That’s why we take the security of your business seriously. When you’re under attack, we’ll work quickly to effectively mitigate attacks and vulnerabilities, and get you back up and running.So next time you are under security emergency please contact F5 SIRT. Processor architecture back in the news First up; processor architecture and vulnerabilities are back in the news again! In the last week we have seen at least three new processor vulnerabilities disclosed - the first two, GhostRace and IPI, both impact all (or most, at least) processor architectures; x86, ARM, RISC-V etc, and deal with exploiting speculative race conditions. In other words, they are a lot like Spectre & Meltdown (there's a great lightboard session if you need a reminder of that research from 2018). GhostRace (CVE-2024-2193) deals with bypassing of synchronization primitives - that is, code designed to prevent multiple threads from trampling on a common resource - like mutexes and spinlocks by using a Spectre-v1 attack on the speculative execution paths. The research was undertaken specifically on Linux although it is quite likely (at least in my opinion) that other operating systems will be similarly affected, but at least on Linux the research showed that the Linux kernel itself is vulnerable to this style of attack. IPI Storming (CVE-2024-26602) is the second half of the complete attack demonstrated in the paper(s); because successful exploitation requires the attacker to win a race condition, exploitation would be most successful if winning that race could be guaranteed. That's where IPI Storming comes in. IPI stands for Inter-Process Interrupt, and the researchers found they could indefinitely 'pause' the victim process by flooding the CPU core it is executing on with interrupts. This allows the attacker to pause the victim process while it is within it's 'race' window (how much of a race is it if the other competitor isn't moving, though) and carry out an infinite number of Speculative Concurrent Use-After-Free attempts. Once successful this attack can leak memory from a target process or, indeed, the Linux kernel at a rate of 12KB/s. As with Spectre, Meltdown and most other hardware-level vulnerabilities that we have seen, the only real mitigation here is to ensure you never run untrusted code on your systems. While there is partial mitigation being added to the Linux kernel for CVE-2024-26602, completely mitigating these issues would reduce processor and operating system performance so significantly as to be improbable. Most (if not all) processor vendors have responded that the mitigations for GhostRace are idenitcal to those for Spectre v1, in fact. That said, the researchers suggest only a 5% performance overhead for their suggested mitigations, albeit in synthetic benchmarks. Despite that assertion I don't believe we'll see widespread adoption of kernel-level or microarchitecture level fixes for this or any other similar vulnerabilities; simply put, we are too addicted to the performance boost speculative execution gives us in modern processors. Just in case Apple were feeling left out with the focus here being on x86 and Linux, we also saw GoFetchpublished last week. This flaw is specific to Apple's M1, M2 and M3 silicon and can allow an attacker access to cryptographic keys, allowing them to access encrypted files or any other data protected by those keys. Once again the research deals with exploiting speculative predictions - in this case we aren't predicting branches in code (as per Spectre) but rather the processor's Data Memory-dependent Prefetcher (DMP) tries to predict which memory address the running code is going to request next so that it can be pre-fected into cache to improve performance. Apparently sometimes, however, the DMP gets the addresses mixed up with thecontents of those addresses, and the attacker can then use cache side channels to guess parts of a cryptograhic key until they've recovered the entire key, and the entire attack can be carried out by unprivileged code running as a regular user. All of the suggested fixes impact performance in some way (did I mention we are addicted to performance, already?) though they have suggested mitigations that might be acceptable to the average user such as pinning cryptographic functions to the "efficiency" cores (which lack DMP functionality) or disable DMP when running security critical applications, leaving it available for performance-focused applications. But, as with the other issues we've talked about today, what manufacturer wants to reduce performance when performance benchmarks are one of the very things they rely on to differentiate themselves from the competition - especially for Apple, who leaned heavily into their roots of media production with the M-silicon and saw huge performance gains (talking to AubreyKingF5 , This Month in Security editor and vidographer extraordinaire, I know that a 4k video which previously took 4 hours to render on his 12 core Intel Macbook now takes 20 minutes thanks to the M-silicon - or, as he put it: "I used to have to set my videos to render overnight. Now, I go get a cup of decaf and it’s done when I’m back."). Would they really want to give any of that up, or push the responsibility back to the user under the guise of "don't run untrusted code"? https://www.vusec.net/projects/ghostrace/ https://www.macworld.com/article/2276399/gofetch-flaw-m1-m2-m3-processor-cryptographic-performance.html Packet goes in, packet goes out, packet goes in, packet goes out.. If you've worked in the networking space for any amount of time, you've probably created a routing or bridging loop at least once in your career, right? Just me? Oh.. I jest, but at least when trying to reproduce parts of complex environments in the lab, it is (in my experience, anyway!) rather easy to accidentally create a loop and have no spanning tree or other mitigations in place and you can quickly take down an entire switching infrastructure - honestly, I have done it more than once although it has, thankfully, been a while. We see them in the F5 SIRT, too; sometimes what presents as a world-ending Distributed Denial of Service (DDoS) turns out to actually be a loop between devices or within a single device, and that storm of traffic was simply a device talking back to itself as fast as it possibly could. Well, enter LoopDOS. I'm going to say something perhaps contraversial here; LoopDOS should not come as a surprise to any of us. At it's core, LoopDOS simply states that (I'm paraphrasing) "The UDP protocol lacks any protection against IP spoofing, this allows an attacker to spoof the source IP on malicious traffic directed to Victim A and have the responses sent to Victim B" and this isnot really news. In fact that is the core of many well understood (and well used!) attacks like DNS amplification where a small request from the attacker sends a large response from a target DNS server back to the ultimate victim; what's new in LoopDOS, however, is that the researchers have identified a number of UDP-based services which can be exploited such that a single packet from the attacker will cause Victim A and Victim B to talk to each otherforever,importantly, the vulnerable services included some very common services like DNS and NTP (plus TFTP, though I think it is fair to say that TFTP is less widely exposed to the Internet than DNS and NTP!). They also demonstrated that legacy services such as QOTD, Chargen, Echo, Time, Daytime and Active Users were vulnerable to the same attack and linked to an old advisory from 1996, CA-1996-01 and honestly this is where I start to have a problem.. CA-1996-01 describes exactly what LoopDOS describes - and it described it28 years ago. The advice in CA-1996-01 and LoopDOS is also largely the same - be very careful when you expose UDP services, don't expose legacy UDP services like chargen and echo (also widely documented as amplification vectors already) monitor traffic and implement rate limiting mechanisms to prevent a complete denial-of-service should a loop occur (the LoopDOS authors also recommend migrating to TCP where possible); so AFM's L4DOS protections are your friend, here. I'm not discounting the usefulness or validity of the researcher, but giving it a catchy name and all the press coverage it's received seemstoo much for what is a fundamental part of a protocl which has existed for 40 years, and describing a flaw first described nearly 30 years prior. Still, here I am, adding to the news coverage.. https://cispa.de/en/loop-dos https://vuls.cert.org/confluence/display/historical/CERT+Advisory+CA-1996-01+UDP+Port+Denial-of-Service+Attack AI worms No, not the ones in the video game, but the kind that infiltrate computer systems and do damage or steal data. Let's talk about ComPromptMized! My first thought was that this was another piece of research on using generative AI to create malware but no - in fact, this is malware that targets generative AI powered applications like email assistants to coerce those assistance into taking malicious action like sending spam or exfiltrating the legitimate user's data. This is far more intriguing to me than the news we've seen in past months about GenAI writing polymorphic malware; this is bascially automated AI prompt-engineering which could be done at-scale to quietly subvert existing infrastructure (such as email) to very quietly exfiltrate data. That's much harder to detect than a new piece of malware trampling across systems (IMHO) as the exfil channel is legitimate, expected, activity. Once again the only mitigation I can really think of is tightly controlling what software is allowed to run on endpoint devices - ensure users can't download arbitrary executables, white-list only known approved software and so on (of course, antivirus/antimalware/etc tooling has a role to play, but that is shown to be routinely bypassed). As a user, I don't really want to see draconian lockdowns on devices but, as a security professional, I can feel it coming.. https://sites.google.com/view/compromptmized More Bad Bots While we are on the topic of malware, I'd like to reference an article that is from slightly outside of "my week" and point everyone to F5 Labs' 2024 Bad Bots Review published March 14th. This is a great drill down into the kinds and amounts of bot traffic mobile APIs and web applications are seeing in the real world, trends across industry verticals (healthcare and hospitality being the leaders in the charts for web traffic, while entertainment leads the way for mobile APIs) as well as across functionality (e.g., search & giftcards are most often automated). It's a 15 minute read and, in my opinion, well worth giving the time to! Of course if (and when) you find yourself on the receiving end of bot traffic, as most sites will at some point, F5 Distributed Cloud Bot Defense can help, as can the F5 SIRTs Emergency Security Response Process if you would rather mitigate using on premise solutions! https://www.f5.com/labs/articles/threat-intelligence/2024-bad-bots-reviewCipher Suite Practices and Pitfalls
Cipher Suite Practices and Pitfalls It seems like every time you turn around there is a new vulnerability to deal with, and some of them, such as Sweet32, have required altering cipher configurations for mitigation. Still other users may tweak their cipher suite settings to meet requirements for PCI compliance, regulatory issues, local compatibility needs, etc. However, once you start modifying your cipher suite settings you must take great care, as it is very easy to shoot yourself in the foot. Many misconfigurations will silently fail – seeming to achieve the intended result while opening up new, even worse, vulnerabilities. Let's take a look at cipher configuration on the F5 BIG-IP products to try stay on the safe path. What is a Cipher Suite? Before we talk about how they're configured, let's define exactly what we mean by 'cipher suite', how it differs from just a 'cipher', and the components of the suite. Wikipedia had a good summary, so rather than reinvent the wheel: A cipher suite is a named combination of authentication, encryption, message authentication code (MAC) and key exchange algorithms used to negotiate the security settings for a network connection using the Transport Layer Security (TLS) / Secure Sockets Layer (SSL) network protocol. When we talk about configuring ciphers on BIG-IP we're really talking about configuring cipher suites. More specifically the configured list of cipher suites is a menu of options available to be negotiated. Each cipher suite specifies the key exchange algorithm, authentication algorithm, cipher, cipher mode, and MAC that will be used. I recommend reading K15194: Overview of the BIG-IP SSL/TLS cipher suite for more information. But as a quick overview, let's look at a couple of example cipher suites. The cipher suite is in the format: Key Exchange-Authentication-Cipher-Cipher Mode-MAC Note that not all of these components may be explicitly present in the cipher suite, but they are still implicitly part of the suite. Let's consider this cipher suite: ECDHE-RSA-AES256-GCM-SHA384 This breaks down as follows: Key Exchange Algorithm: ECDHE (Elliptic Curve Diffie-Hellman Ephemeral) Authentication Algorithm: RSA Cipher: AES256 (aka AES with a 256-bit key) Cipher Mode: GCM (Galois/Counter Mode) MAC: SHA384 (aka SHA-2 (Secure Hash Algorithm 2) with 384-bit hash) This is arguably the strongest cipher suite we have on BIG-IP at this time. Let's compare that to a simpler cipher suite: AES128-SHA Key Exchange Algorithm: RSA (Implied) – When it isn't specified, presume RSA. Authentication Algorithm: RSA (Implied) – When it isn't specified, presume RSA. Cipher: AES128 (aka AES with a 128-bit key) Cipher Mode: CBC (Cipher Block Chaining) (Implied) – When it isn't specified, presume CBC. MAC: SHA1 (Secure Hash Algorithm 1; SHA-1 always produces a 160-bit hash.) This example illustrates that the cipher suite may not always explicitly specify every parameter, but they're still there. There are 'default' values that are fairly safe to presume when not otherwise specified. If an algorithm isn't specified, it is RSA. That's a safe bet. And if a cipher mode isn't specified it is CBC. Always CBC. Note that all ciphers currently supported on BIG-IP are CBC mode except for AES-GCM and RC4. ALL. I stress this as it has been a recurring source of confusion amongst customers. It isn't only the cipher suites which explicitly state 'CBC' in their name. Let's examine each of these components. This article is primarily about cipher suite configuration and ciphers, and not the SSL/TLS protocol, so I won't dive too deeply here, but I think it helps to have a basic understanding. Forgive me if I simplify a bit. Key Exchange Algorithms As a quick review of the difference between asymmetric key (aka public key) cryptography and symmetric key cryptography: With the asymmetric key you have two keys – K public and K private –which have a mathematical relationship. Since you can openly share the public key there is no need to pre-share keys with anyone. The downside is that these algorithms are computationally expensive. Key lengths for a common algorithm such as RSA are at least 1024-bit, and 2048-bit is really the minimally acceptable these days. Symmetric key has only K private . Both ends use the same key, which poses the problem of key distribution. The advantage is higher computational performance and common key sizes are 128-bit or 256-bit. SSL/TLS, of course, uses both public and private key systems – the Key Exchange Algorithm is the public key system used to exchange the symmetric key. Examples you'll see in cipher suites include ECDHE, DHE, RSA, ECDH, and ADH. Authentication Algorithms The Authentication Algorithm is sometimes grouped in with the Key Exchange Algorithm for configuration purposes; 'ECDHE_RSA' for example. But we'll consider it as a separate component. This is the algorithm used in the SSL/TLS handshake for the server to sign (using the server's private key) elements sent to the client in the negotiation. The client can authenticate them using the server's public key. Examples include: RSA, ECDSA, DSS (aka DSA), and Anonymous. Anonymous means no authentication; this is generally bad. The most common way users run into this is by accidentally enabling an 'ADH' cipher suite. More on this later when we talk about pitfalls. Note that when RSA is used for the key exchange, authentication is inherent to the scheme so there really isn't a separate authentication step. However, most tools will list it out for completeness. Cipher To borrow once again from Wikipedia: In cryptography, a cipher (or cypher) is an algorithm for performing encryption or decryption—a series of well-defined steps that can be followed as a procedure. An alternative, less common term is encipherment. To encipher or encode is to convert information into cipher or code. In common parlance, 'cipher' is synonymous with 'code', as they are both a set of steps that encrypt a message; however, the concepts are distinct in cryptography, especially classical cryptography. This is what most of us mean when we refer to 'configuring ciphers'. We're primarily interested in controlling the cipher used to protect our information through encryption. There are many, many examples of ciphers which you may be familiar with: DES (Data Encryption Standard), 3DES (Triple DES), AES (Advanced Encryption Standard), RC4 (Rivest Cipher 4), Camellia, RC6, RC2, Blowfish, Twofish, IDEA, SEED, GOST, Rijndael, Serpent, MARS, etc. For a little cipher humor, I recommend RFC2410: The NULL Encryption Algorithm and Its Use With IPsec. Roughly speaking, ciphers come in two types – block ciphers and stream ciphers. Block Ciphers Block ciphers operate on fixed-length chunks of data, or blocks. For example, DES operates on 64-bit blocks while AES operates on 128-bit blocks. Most of the ciphers you'll encounter are block ciphers. Examples: DES, 3DES, AES, Blowfish, Twofish, etc. Stream Ciphers Stream ciphers mathematically operate on each bit in the data flow individually. The most commonly encountered stream cipher is RC4, and that's deprecated. So we're generally focused on block ciphers, not that it really changes anything for the purposes of this article. All of the secrecy in encryption comes from the key that is used, not the cipher itself. Obtain the key and you can unlock the ciphertext. The cipher itself – the algorithm, source code, etc. – not only can be, but should be, openly available. History is full of examples of private cryptosystems failing due to weaknesses missed by their creators, while the most trusted ciphers were created via open processes (AES for example). Keys are of varying lengths and, generally speaking, the longer the key the more secure the encryption. DES only had 56-bits of key data, and thus is considered insecure. We label 3DES as 168-bit, but it is really only equivalent to 112-bit strength. (More on this later.) Newer ciphers, such as AES, often offer options – 128-bits, 192-bits, or 256-bits of key. Remember, a 256-bit key is far more than twice as strong as a 128-bit key. It is 2 128 vs. 2 256 - 3.4028237e+38 vs. 1.1579209e+77 Cipher Mode Cipher mode is the mode of operation used by the cipher when encrypting plaintext into ciphertext, or decrypting ciphertext into plaintext. The most common mode is CBC – Cipher Block Chaining. In cipher block chaining the ciphertext from block n feeds into the process for block n+1 – the blocks are chained together. To steal borrow an image from Wikipedia: As I mentioned previously, all ciphers on BIG-IP are CBC mode except for RC4 (the lone stream cipher, disabled by default starting in 11.6.0) and AES-GCM. AES-GCM was first introduced in 11.5.0, and it is only available for TLSv1.2 connections. GCM stands for Galois/Counter Mode, a more advanced mode of operation than CBC. In GCM the blocks are not chained together. GCM runs in an Authenticated Encryption with Associated Data (AEAD) mode which eliminates the separate per-message hashing step, therefore it can achieve higher performance than CBC mode on a given HW platform. It is also immune to classes of attack that have harried CBC, such as the numerous padding attacks (BEAST, Lucky 13, etc.) Via Wikipedia: The main drawback to AES-GCM is that it was only added in TLSv1.2, so any older clients which don't support TLSv1.2 cannot use it. There are other cipher suites officially supported in TLS which have other modes, but F5 does not currently support those ciphers so we won't get too deep into that. Other ciphers include AES-CCM (CTR mode with a CBC MAC; CTR is Counter Mode), CAMELLIA-GCM (CAMELLIA as introduced in 12.0.0 is CBC), and GOST CNT (aka CTR). We may see these in the future. MAC aka Hash Function What did we ever do before Wikipedia? A hash function is any function that can be used to map data of arbitrary size to data of fixed size. The values returned by a hash function are called hash values, hash codes, digests, or simply hashes. One use is a data structure called a hash table, widely used in computer software for rapid data lookup. Hash functions accelerate table or database lookup by detecting duplicated records in a large file. An example is finding similar stretches in DNA sequences. They are also useful in cryptography. A cryptographic hash function allows one to easily verify that some input data maps to a given hash value, but if the input data is unknown, it is deliberately difficult to reconstruct it (or equivalent alternatives) by knowing the stored hash value. This is used for assuring integrity of transmitted data, and is the building block for HMACs, which provide message authentication. In short, the MAC provides message integrity. Hash functions include MD5, SHA-1 (aka SHA), SHA-2 (aka SHA128, SHA256, & SHA384), and AEAD (Authenticated Encryption with Associated Data). MD5 has long since been rendered completely insecure and is deprecated. SHA-1 is now being 'shamed', if not blocked, by browsers as it is falling victim to advances in cryptographic attacks. While some may need to continue to support SHA-1 cipher suites for legacy clients, it is encouraged to migrate to SHA-2 as soon as possible – especially for digital certificates. Configuring Cipher Suites on BIG-IP Now that we've covered what cipher suites are, let's look at where we use them. There are two distinct and separate areas where cipher suites are used – the host, or control plane, and TMM, or the data plane. On the host side SSL/TLS is handled by OpenSSL and the configuration follows the standard OpenSSL configuration options. Control Plane The primary use of SSL/TLS on the control plane is for httpd. To see the currently configured cipher suite, use ' tmsh list sys http ssl-ciphersuite '. The defaults may vary depending on the version of TMOS. For example, these were the defaults in 12.0.0: tmsh list sys http ssl-ciphersuite sys httpd { ssl-ciphersuite DEFAULT:!aNULL:!eNULL:!LOW:!RC4:!MD5:!EXP } As of 12.1.2 these have been updated to a more explicit list: tmsh list sys http ssl-ciphersuite sys httpd { ssl-ciphersuite ECDHE-RSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-SHA:ECDHE-RSA-AES256-SHA:ECDHE-RSA-AES128-SHA256:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES128-SHA:ECDHE-ECDSA-AES256-SHA:ECDHE-ECDSA-AES128-SHA256:ECDHE-ECDSA-AES256-SHA384:AES128-GCM-SHA256:AES256-GCM-SHA384:AES128-SHA:AES256-SHA:AES128-SHA256:AES256-SHA256:ECDHE-RSA-DES-CBC3-SHA:ECDHE-ECDSA-DES-CBC3-SHA:DES-CBC3-SHA } You can change this configuration via ' tmsh modify sys http ssl-ciphersuite <value> '. One important thing to note is that the default is not just 'DEFAULT' as it is on the data plane. This is one thing that users have been caught by; thinking that setting the keyword to 'DEFAULT' will reset the configuration. As OpenSSL provides SSL/TLS support for the control plane, if you want to see which ciphers will actually be supported you can use ' openssl ciphers -v <cipherstring> '. For example: openssl ciphers -v 'ECDHE-RSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-SHA:ECDHE-RSA-AES256-SHA:ECDHE-RSA-AES128-SHA256:ECDHE-RSA-AES256-SHA384:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-ECDSA-AES128-SHA:ECDHE-ECDSA-AES256-SHA:ECDHE-ECDSA-AES128-SHA256:ECDHE-ECDSA-AES256-SHA384:AES128-GCM-SHA256:AES256-GCM-SHA384:AES128-SHA:AES256-SHA:AES128-SHA256:AES256-SHA256:ECDHE-RSA-DES-CBC3-SHA:ECDHE-ECDSA-DES-CBC3-SHA:DES-CBC3-SHA' ECDHE-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(128) Mac=AEAD ECDHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(256) Mac=AEAD ECDHE-RSA-AES128-SHA SSLv3 Kx=ECDH Au=RSA Enc=AES(128) Mac=SHA1 ECDHE-RSA-AES256-SHA SSLv3 Kx=ECDH Au=RSA Enc=AES(256) Mac=SHA1 ECDHE-RSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(128) Mac=SHA256 ECDHE-RSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(256) Mac=SHA384 ECDHE-ECDSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(128) Mac=AEAD ECDHE-ECDSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(256) Mac=AEAD ECDHE-ECDSA-AES128-SHA SSLv3 Kx=ECDH Au=ECDSA Enc=AES(128) Mac=SHA1 ECDHE-ECDSA-AES256-SHA SSLv3 Kx=ECDH Au=ECDSA Enc=AES(256) Mac=SHA1 ECDHE-ECDSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(128) Mac=SHA256 ECDHE-ECDSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(256) Mac=SHA384 AES128-GCM-SHA256 TLSv1.2 Kx=RSA Au=RSA Enc=AESGCM(128) Mac=AEAD AES256-GCM-SHA384 TLSv1.2 Kx=RSA Au=RSA Enc=AESGCM(256) Mac=AEAD AES128-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(128) Mac=SHA1 AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1 AES128-SHA256 TLSv1.2 Kx=RSA Au=RSA Enc=AES(128) Mac=SHA256 AES256-SHA256 TLSv1.2 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA256 ECDHE-RSA-DES-CBC3-SHA SSLv3 Kx=ECDH Au=RSA Enc=3DES(168) Mac=SHA1 ECDHE-ECDSA-DES-CBC3-SHA SSLv3 Kx=ECDH Au=ECDSA Enc=3DES(168) Mac=SHA1 DES-CBC3-SHA SSLv3 Kx=RSA Au=RSA Enc=3DES(168) Mac=SHA1 Now let's see what happens if you use 'DEFAULT': openssl ciphers -v 'DEFAULT' ECDHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(256) Mac=AEAD ECDHE-ECDSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(256) Mac=AEAD ECDHE-RSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(256) Mac=SHA384 ECDHE-ECDSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(256) Mac=SHA384 ECDHE-RSA-AES256-SHA SSLv3 Kx=ECDH Au=RSA Enc=AES(256) Mac=SHA1 ECDHE-ECDSA-AES256-SHA SSLv3 Kx=ECDH Au=ECDSA Enc=AES(256) Mac=SHA1 DHE-DSS-AES256-GCM-SHA384 TLSv1.2 Kx=DH Au=DSS Enc=AESGCM(256) Mac=AEAD DHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=DH Au=RSA Enc=AESGCM(256) Mac=AEAD DHE-RSA-AES256-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AES(256) Mac=SHA256 DHE-DSS-AES256-SHA256 TLSv1.2 Kx=DH Au=DSS Enc=AES(256) Mac=SHA256 DHE-RSA-AES256-SHA SSLv3 Kx=DH Au=RSA Enc=AES(256) Mac=SHA1 DHE-DSS-AES256-SHA SSLv3 Kx=DH Au=DSS Enc=AES(256) Mac=SHA1 DHE-RSA-CAMELLIA256-SHA SSLv3 Kx=DH Au=RSA Enc=Camellia(256) Mac=SHA1 DHE-DSS-CAMELLIA256-SHA SSLv3 Kx=DH Au=DSS Enc=Camellia(256) Mac=SHA1 ECDH-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH/RSA Au=ECDH Enc=AESGCM(256) Mac=AEAD ECDH-ECDSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH/ECDSA Au=ECDH Enc=AESGCM(256) Mac=AEAD ECDH-RSA-AES256-SHA384 TLSv1.2 Kx=ECDH/RSA Au=ECDH Enc=AES(256) Mac=SHA384 ECDH-ECDSA-AES256-SHA384 TLSv1.2 Kx=ECDH/ECDSA Au=ECDH Enc=AES(256) Mac=SHA384 ECDH-RSA-AES256-SHA SSLv3 Kx=ECDH/RSA Au=ECDH Enc=AES(256) Mac=SHA1 ECDH-ECDSA-AES256-SHA SSLv3 Kx=ECDH/ECDSA Au=ECDH Enc=AES(256) Mac=SHA1 AES256-GCM-SHA384 TLSv1.2 Kx=RSA Au=RSA Enc=AESGCM(256) Mac=AEAD AES256-SHA256 TLSv1.2 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA256 AES256-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(256) Mac=SHA1 CAMELLIA256-SHA SSLv3 Kx=RSA Au=RSA Enc=Camellia(256) Mac=SHA1 PSK-AES256-CBC-SHA SSLv3 Kx=PSK Au=PSK Enc=AES(256) Mac=SHA1 ECDHE-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(128) Mac=AEAD ECDHE-ECDSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(128) Mac=AEAD ECDHE-RSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(128) Mac=SHA256 ECDHE-ECDSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(128) Mac=SHA256 ECDHE-RSA-AES128-SHA SSLv3 Kx=ECDH Au=RSA Enc=AES(128) Mac=SHA1 ECDHE-ECDSA-AES128-SHA SSLv3 Kx=ECDH Au=ECDSA Enc=AES(128) Mac=SHA1 DHE-DSS-AES128-GCM-SHA256 TLSv1.2 Kx=DH Au=DSS Enc=AESGCM(128) Mac=AEAD DHE-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AESGCM(128) Mac=AEAD DHE-RSA-AES128-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AES(128) Mac=SHA256 DHE-DSS-AES128-SHA256 TLSv1.2 Kx=DH Au=DSS Enc=AES(128) Mac=SHA256 DHE-RSA-AES128-SHA SSLv3 Kx=DH Au=RSA Enc=AES(128) Mac=SHA1 DHE-DSS-AES128-SHA SSLv3 Kx=DH Au=DSS Enc=AES(128) Mac=SHA1 DHE-RSA-SEED-SHA SSLv3 Kx=DH Au=RSA Enc=SEED(128) Mac=SHA1 DHE-DSS-SEED-SHA SSLv3 Kx=DH Au=DSS Enc=SEED(128) Mac=SHA1 DHE-RSA-CAMELLIA128-SHA SSLv3 Kx=DH Au=RSA Enc=Camellia(128) Mac=SHA1 DHE-DSS-CAMELLIA128-SHA SSLv3 Kx=DH Au=DSS Enc=Camellia(128) Mac=SHA1 ECDH-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH/RSA Au=ECDH Enc=AESGCM(128) Mac=AEAD ECDH-ECDSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH/ECDSA Au=ECDH Enc=AESGCM(128) Mac=AEAD ECDH-RSA-AES128-SHA256 TLSv1.2 Kx=ECDH/RSA Au=ECDH Enc=AES(128) Mac=SHA256 ECDH-ECDSA-AES128-SHA256 TLSv1.2 Kx=ECDH/ECDSA Au=ECDH Enc=AES(128) Mac=SHA256 ECDH-RSA-AES128-SHA SSLv3 Kx=ECDH/RSA Au=ECDH Enc=AES(128) Mac=SHA1 ECDH-ECDSA-AES128-SHA SSLv3 Kx=ECDH/ECDSA Au=ECDH Enc=AES(128) Mac=SHA1 AES128-GCM-SHA256 TLSv1.2 Kx=RSA Au=RSA Enc=AESGCM(128) Mac=AEAD AES128-SHA256 TLSv1.2 Kx=RSA Au=RSA Enc=AES(128) Mac=SHA256 AES128-SHA SSLv3 Kx=RSA Au=RSA Enc=AES(128) Mac=SHA1 SEED-SHA SSLv3 Kx=RSA Au=RSA Enc=SEED(128) Mac=SHA1 CAMELLIA128-SHA SSLv3 Kx=RSA Au=RSA Enc=Camellia(128) Mac=SHA1 PSK-AES128-CBC-SHA SSLv3 Kx=PSK Au=PSK Enc=AES(128) Mac=SHA1 ECDHE-RSA-RC4-SHA SSLv3 Kx=ECDH Au=RSA Enc=RC4(128) Mac=SHA1 ECDHE-ECDSA-RC4-SHA SSLv3 Kx=ECDH Au=ECDSA Enc=RC4(128) Mac=SHA1 ECDH-RSA-RC4-SHA SSLv3 Kx=ECDH/RSA Au=ECDH Enc=RC4(128) Mac=SHA1 ECDH-ECDSA-RC4-SHA SSLv3 Kx=ECDH/ECDSA Au=ECDH Enc=RC4(128) Mac=SHA1 RC4-SHA SSLv3 Kx=RSA Au=RSA Enc=RC4(128) Mac=SHA1 RC4-MD5 SSLv3 Kx=RSA Au=RSA Enc=RC4(128) Mac=MD5 PSK-RC4-SHA SSLv3 Kx=PSK Au=PSK Enc=RC4(128) Mac=SHA1 ECDHE-RSA-DES-CBC3-SHA SSLv3 Kx=ECDH Au=RSA Enc=3DES(168) Mac=SHA1 ECDHE-ECDSA-DES-CBC3-SHA SSLv3 Kx=ECDH Au=ECDSA Enc=3DES(168) Mac=SHA1 EDH-RSA-DES-CBC3-SHA SSLv3 Kx=DH Au=RSA Enc=3DES(168) Mac=SHA1 EDH-DSS-DES-CBC3-SHA SSLv3 Kx=DH Au=DSS Enc=3DES(168) Mac=SHA1 ECDH-RSA-DES-CBC3-SHA SSLv3 Kx=ECDH/RSA Au=ECDH Enc=3DES(168) Mac=SHA1 ECDH-ECDSA-DES-CBC3-SHA SSLv3 Kx=ECDH/ECDSA Au=ECDH Enc=3DES(168) Mac=SHA1 DES-CBC3-SHA SSLv3 Kx=RSA Au=RSA Enc=3DES(168) Mac=SHA1 PSK-3DES-EDE-CBC-SHA SSLv3 Kx=PSK Au=PSK Enc=3DES(168) Mac=SHA1 EDH-RSA-DES-CBC-SHA SSLv3 Kx=DH Au=RSA Enc=DES(56) Mac=SHA1 EDH-DSS-DES-CBC-SHA SSLv3 Kx=DH Au=DSS Enc=DES(56) Mac=SHA1 DES-CBC-SHA SSLv3 Kx=RSA Au=RSA Enc=DES(56) Mac=SHA1 EXP-EDH-RSA-DES-CBC-SHA SSLv3 Kx=DH(512) Au=RSA Enc=DES(40) Mac=SHA1 export EXP-EDH-DSS-DES-CBC-SHA SSLv3 Kx=DH(512) Au=DSS Enc=DES(40) Mac=SHA1 export EXP-DES-CBC-SHA SSLv3 Kx=RSA(512) Au=RSA Enc=DES(40) Mac=SHA1 export EXP-RC2-CBC-MD5 SSLv3 Kx=RSA(512) Au=RSA Enc=RC2(40) Mac=MD5 export EXP-RC4-MD5 SSLv3 Kx=RSA(512) Au=RSA Enc=RC4(40) Mac=MD5 export As you can see that enables far, far more ciphers, including a number of unsafe ciphers – export, MD5, DES, etc. This is a good example of why you always want to confirm your cipher settings and check exactly what is being enabled before placing new settings into production. Many security disasters could be avoided if everyone doublechecked their settings first. Let’s take a closer look at how OpenSSL represents one of the cipher suites: ECDHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(256) Mac=AEAD The columns are: Cipher Suite: ECDHE-RSA-AES256-GCM-SHA384 Protocol: TLSv1.2 Key Exchange Algorithm (Kx): ECDH Authentication Algorithm (Au): RSA Cipher/Encryption Algorithm (Enc): AESGCM(256) MAC (Mac): AEAD Since the control plane uses OpenSSL you can use the standard OpenSSL documentation, so I won't spend a lot of time on that. Data Plane In TMM the cipher suites are configured in the Ciphers field of the Client SSL or Server SSL profiles. See K14783: Overview of the Client SSL profile (11.x - 12.x) & K14806: Overview of the Server SSL profile (11.x - 12.x), respectively for more details. It is important to keep in mind that these are two different worlds with their own requirements and quirks. As most of the configuration activity, and security concerns, occur on the public facing side of the system, we'll focus on the Client SSL Profile. Most of the things we'll cover here will also apply to the Server SSL profile. In the GUI it appears as an editable field: Presuming the profile was created with the name 'Test': tmsh list ltm profile client-ssl Test ltm profile client-ssl Test { app-service none cert default.crt cert-key-chain { default { cert default.crt key default.key } } chain none ciphers DEFAULT defaults-from clientssl inherit-certkeychain true key default.key passphrase none } Modifying the cipher configuration from the command line is simple. tmsh list ltm profile client-ssl Test ciphers ltm profile client-ssl Test { ciphers DEFAULT } tmsh modify ltm profile client-ssl Test ciphers 'DEFAULT:!3DES' tmsh list ltm profile client-ssl Test ciphers ltm profile client-ssl Test { ciphers DEFAULT:!3DES } Just remember the ' tmsh save sys config ' when you're happy with the configuration. Note here the default is just 'DEFAULT'. What that expands to will vary depending on the version of TMOS. K13156: SSL ciphers used in the default SSL profiles (11.x - 12.x) defines the default values for each version of TMOS. Or you can check it locally from the command line: tmm --clientciphers 'DEFAULT' On 12.1.2 that would be: tmm --clientciphers 'DEFAULT' ID SUITE BITS PROT METHOD CIPHER MAC KEYX 0: 159 DHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 EDH/RSA 1: 158 DHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 EDH/RSA 2: 107 DHE-RSA-AES256-SHA256 256 TLS1.2 Native AES SHA256 EDH/RSA 3: 57 DHE-RSA-AES256-SHA 256 TLS1 Native AES SHA EDH/RSA 4: 57 DHE-RSA-AES256-SHA 256 TLS1.1 Native AES SHA EDH/RSA 5: 57 DHE-RSA-AES256-SHA 256 TLS1.2 Native AES SHA EDH/RSA 6: 57 DHE-RSA-AES256-SHA 256 DTLS1 Native AES SHA EDH/RSA 7: 103 DHE-RSA-AES128-SHA256 128 TLS1.2 Native AES SHA256 EDH/RSA 8: 51 DHE-RSA-AES128-SHA 128 TLS1 Native AES SHA EDH/RSA 9: 51 DHE-RSA-AES128-SHA 128 TLS1.1 Native AES SHA EDH/RSA 10: 51 DHE-RSA-AES128-SHA 128 TLS1.2 Native AES SHA EDH/RSA 11: 51 DHE-RSA-AES128-SHA 128 DTLS1 Native AES SHA EDH/RSA 12: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1 Native DES SHA EDH/RSA 13: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1.1 Native DES SHA EDH/RSA 14: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1.2 Native DES SHA EDH/RSA 15: 22 DHE-RSA-DES-CBC3-SHA 168 DTLS1 Native DES SHA EDH/RSA 16: 157 AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 RSA 17: 156 AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 RSA 18: 61 AES256-SHA256 256 TLS1.2 Native AES SHA256 RSA 19: 53 AES256-SHA 256 TLS1 Native AES SHA RSA 20: 53 AES256-SHA 256 TLS1.1 Native AES SHA RSA 21: 53 AES256-SHA 256 TLS1.2 Native AES SHA RSA 22: 53 AES256-SHA 256 DTLS1 Native AES SHA RSA 23: 60 AES128-SHA256 128 TLS1.2 Native AES SHA256 RSA 24: 47 AES128-SHA 128 TLS1 Native AES SHA RSA 25: 47 AES128-SHA 128 TLS1.1 Native AES SHA RSA 26: 47 AES128-SHA 128 TLS1.2 Native AES SHA RSA 27: 47 AES128-SHA 128 DTLS1 Native AES SHA RSA 28: 10 DES-CBC3-SHA 168 TLS1 Native DES SHA RSA 29: 10 DES-CBC3-SHA 168 TLS1.1 Native DES SHA RSA 30: 10 DES-CBC3-SHA 168 TLS1.2 Native DES SHA RSA 31: 10 DES-CBC3-SHA 168 DTLS1 Native DES SHA RSA 32: 49200 ECDHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 ECDHE_RSA 33: 49199 ECDHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 ECDHE_RSA 34: 49192 ECDHE-RSA-AES256-SHA384 256 TLS1.2 Native AES SHA384 ECDHE_RSA 35: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1 Native AES SHA ECDHE_RSA 36: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1.1 Native AES SHA ECDHE_RSA 37: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1.2 Native AES SHA ECDHE_RSA 38: 49191 ECDHE-RSA-AES128-SHA256 128 TLS1.2 Native AES SHA256 ECDHE_RSA 39: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1 Native AES SHA ECDHE_RSA 40: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1.1 Native AES SHA ECDHE_RSA 41: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1.2 Native AES SHA ECDHE_RSA 42: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1 Native DES SHA ECDHE_RSA 43: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1.1 Native DES SHA ECDHE_RSA 44: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1.2 Native DES SHA ECDHE_RSA Some differences when compared to OpenSSL are readily apparent. For starters, TMM kindly includes a column label header, and actually aligns the columns. The first column is simply a 0-ordinal numeric index, the rest are as follows: ID: The official SSL/TLS ID assigned to that cipher suite. SUITE: The cipher suite. BITS: The size of the key in bits. PROT: The protocol supported. METHOD: NATIVE (in TMM) vs. COMPAT (using OpenSSL code). CIPHER: The cipher. MAC: The hash function. KEYX: The Key Exchange and Authentication Algorithms Note that the MAC is a little misleading for AES-GCM cipher suites. There is no separate MAC as they're AEAD. But the hashing algorithm is used in the Pseudo-Random Function (PRF) and a few other handshake related places. Selecting the Cipher Suites Now we know how to look at the current configuration, modify it, and list the actual ciphers that will be enabled by the listed suites. But what do we put into the configuration? Most users won't have to touch this. The default values are carefully selected by F5 to meet the needs of the majority of our customers. That's the good news. The bad news is that some customers will need to get in there and change the configuration – be it for regulatory compliance, internal policies, legacy client support, etc. Once you begin modifying them, the configuration is truly custom for each customer. Every customer who modifies the configuration, and uses a custom cipher configuration, needs to determine what the proper list is for their needs. Let's say we have determined that we need to support only AES & AES-GCM, 128-bit or 256-bit, and only ECDHE key exchange. Any MAC or Authentication is fine. OK, let's proceed from there. On 12.1.2 there are six cipher suites that fit those criteria. We could list them all explicitly: tmm --clientciphers 'ECDHE-RSA-AES256-GCM-SHA384:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-RSA-AES256-SHA384:ECDHE-RSA-AES256-CBC-SHA:ECDHE-RSA-AES128-SHA256:ECDHE-RSA-AES128-CBC-SHA' That will work, but it gets unwieldy fast. Not only that, but in versions up to 11.5.0 the ciphers configuration string was truncated at 256bytes. Starting in 11.5.0 that was increased to 768bytes, but that can still truncate long configurations. We'll revisit this when we get to the pitfalls section. Fortunately, there is an alternative – keywords! This will result in the same list of cipher suites: tmm --clientciphers 'ECDHE+AES-GCM:ECDHE+AES' That specifies the ECDHE key exchange with AES-GCM ciphers, and ECDHE with AES ciphers. Let's take a closer look to help understand what is happening here. Keywords Keywords are extremely important when working with cipher suite configuration, so we'll spend a little time on those. Most of these apply to both the control plane (OpenSSL) and the data plane (TMM), unless otherwise noted, but we're focused on the data plane as that's F5 specific. Keywords organize into different categories. F5 specific: NATIVE: cipher suites implemented natively in TMM COMPAT: cipher suites using OpenSSL code; removed as of 12.0.0 @SPEED: Re-orders the list to put 'faster' (based on TMOS implementation performance) ciphers first. Sorting: @SPEED: Re-orders the list to put 'faster' (based on TMOS implementation performance) ciphers first. (F5 Specific) @STRENGTH: Re-orders the list to put 'stronger' (larger keys) ciphers first. Protocol: TLSv1_2: cipher suites available under TLSv1.2 TLSv1_1: cipher suites available under TLSv1.1 TLSv1: cipher suites available under TLSv1.0 SSLv3: cipher suites available under SSLv3 Note the 'Protocol' keywords in the cipher configuration control the ciphers associated with that protocol, and not the protocol itself! More on this in pitfalls. Key Exchange Algorithms (sometimes with Authentication specified): ECDHE or ECDHA_RSA: Elliptic Curve Diffie-Hellman Ephemeral (with RSA) ECDHE_ECDSA: ECDHE with Elliptic Curve Digital Signature Algorithm DHE or EDH: Diffie-Hellman Ephemeral (aka Ephemeral Diffie-Hellman) (with RSA) DHE_DSS: DHE with Digital Signature Standard (aka DSA – Digital Signature Algorithm) ECDH_RSA: Elliptic Curve Diffie-Hellman with RSA ECDH_ECDSA: ECDH with ECDSA RSA: RSA, obviously ADH: Anonymous Diffie-Hellman. Note the Authentication Algorithms don't work as standalone keywords in TMM. You can't use 'ECDSA' or 'DSS' for example. And you might think ECDHE or DHE includes all such cipher suites – note that they don't if you read carefully. General cipher groupings: DEFAULT: The default cipher suite for that version; see K13156 ALL: All NATIVE cipher suites; does not include COMPAT in current versions HIGH: 'High' security cipher suites; >128-bit MEDIUM: 'Medium' security cipher suites; effectively 128-bit suites LOW: 'Low' security cipher suites; <128-bit excluding export grade ciphers EXP or EXPORT: Export grade ciphers; 40-bit or 56-bit EXPORT56: 56-bit export ciphers EXPORT40: 40-bit export ciphers Note that DEFAULT does change periodically as F5 updates the configuration to follow the latest best practices. K13156: SSL ciphers used in the default SSL profiles (11.x - 12.x) documents these changes. Cipher families: AES-GCM: AES in GCM mode; 128-bit or 256-bit AES: AES in CBC mode; 128-bit or 256-bit CAMELLIA: Camellia in CBC mode; 128-bit or 256-bit 3DES: Triple DES in CBC mode; 168-bit (well, 112-bit really) DES: Single DES in CBC mode, includes EXPORTciphers;40-bit & 56-bit. RC4: RC4 stream cipher NULL: NULL cipher; just what it sounds like, it does nothing – no encryption MAC aka Hash Function: SHA384: SHA-2 384-bit hash SHA256: SHA-2 256-bit hash SHA1 or SHA: SHA-1 160-bit hash MD5: MD5 128-bit hash Other: On older TMOS versions when using the COMPAT keyword it also enables two additional keywords: SSLv2: Ciphers supported on the SSLv2 protocol RC2: RC2 ciphers. So, let's go back to our example: tmm --clientciphers 'ECDHE+AES-GCM:ECDHE+AES' Note that you can combine keywords using '+' (plus sign). And multiple entries in the ciphers configuration line are separated with ':' (colon). You may also need to wrap the string in single quotes on the command line – I find it is a good habit to just always do so. We can also exclude suites or keywords. There are two ways to do that: '!' (exclamation point) is a hard exclusion. Anything excluded this way cannot be implicitly or explicitly re-enabled. It is disabled, period. '-' (minus sign or dash) is a soft exclusion. Anything excluded this way can be explicitly re-enabled later in the configuration string. (Note: The dash is also usedinthe names of many cipher suites, such as ECDHE-RSA-AES256-GCM-SHA384 or AES128-SHA. Do not confuse the dashes that are part of the cipher suite names with a soft exclusion, which alwaysprecedes, or prefixes,the value being excluded. 'AES128-SHA': AES128-SHA cipher suite. '-SHA': SHA is soft excluded. '-AES128-SHA': the AES128-SHA cipher suite is soft excluded. Position matters.) Let's look at the difference in hard and soft exclusions. We'll start with our base example: tmm --clientciphers 'ECDHE+AES-GCM:DHE+AES-GCM' ID SUITE BITS PROT METHOD CIPHER MAC KEYX 0: 49200 ECDHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 ECDHE_RSA 1: 49199 ECDHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 ECDHE_RSA 2: 159 DHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 EDH/RSA 3: 158 DHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 EDH/RSA Now let's look at a hard exclusion: tmm --clientciphers 'ECDHE+AES-GCM:!DHE:DHE+AES-GCM' ID SUITE BITS PROT METHOD CIPHER MAC KEYX 0: 49200 ECDHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 ECDHE_RSA 1: 49199 ECDHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 ECDHE_RSA And lastly a soft exclusion: tmm --clientciphers 'ECDHE+AES-GCM:-DHE:DHE+AES-GCM' ID SUITE BITS PROT METHOD CIPHER MAC KEYX 0: 49200 ECDHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 ECDHE_RSA 1: 49199 ECDHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 ECDHE_RSA 2: 159 DHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 EDH/RSA 3: 158 DHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 EDH/RSA Note that in the second example, the hard exclusion, we used '!DHE' and even though we then explicitly added 'DHE+AES-GCM' those ciphers were not enabled. This is because, once excluded with a hard exclusion, ciphers cannot be re-enabled. In the third example, the soft exclusion, we used '-DHE' and then 'DHE+AES-GCM'. This time it did enable those ciphers, which is possible with a soft exclusion. You might be wondering what soft disabling is useful for; why would you ever want to remove ciphers only to add them again? Reordering the ciphers is a common use case. As an example, DEFAULT orders ciphers differently in different versions, but mainly based on strength – bit size. Let's say we know 3DES is really 112-bit equivalent strength and not 168-bit as it is usually labeled. For some reason, maybe legacy clients, we can't disable them, but we want them to be last on the list. One way to do this is to first configure the DEFAULT list, then remove all of the 3DES ciphers. But then add the 3DES ciphers back explicitly – at the end of the list. Let's try it – compare the following: tmm --clientciphers 'DEFAULT' ID SUITE BITS PROT METHOD CIPHER MAC KEYX 0: 159 DHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 EDH/RSA 1: 158 DHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 EDH/RSA 2: 107 DHE-RSA-AES256-SHA256 256 TLS1.2 Native AES SHA256 EDH/RSA 3: 57 DHE-RSA-AES256-SHA 256 TLS1 Native AES SHA EDH/RSA 4: 57 DHE-RSA-AES256-SHA 256 TLS1.1 Native AES SHA EDH/RSA 5: 57 DHE-RSA-AES256-SHA 256 TLS1.2 Native AES SHA EDH/RSA 6: 57 DHE-RSA-AES256-SHA 256 DTLS1 Native AES SHA EDH/RSA 7: 103 DHE-RSA-AES128-SHA256 128 TLS1.2 Native AES SHA256 EDH/RSA 8: 51 DHE-RSA-AES128-SHA 128 TLS1 Native AES SHA EDH/RSA 9: 51 DHE-RSA-AES128-SHA 128 TLS1.1 Native AES SHA EDH/RSA 10: 51 DHE-RSA-AES128-SHA 128 TLS1.2 Native AES SHA EDH/RSA 11: 51 DHE-RSA-AES128-SHA 128 DTLS1 Native AES SHA EDH/RSA 12: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1 Native DES SHA EDH/RSA 13: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1.1 Native DES SHA EDH/RSA 14: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1.2 Native DES SHA EDH/RSA 15: 22 DHE-RSA-DES-CBC3-SHA 168 DTLS1 Native DES SHA EDH/RSA 16: 157 AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 RSA 17: 156 AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 RSA 18: 61 AES256-SHA256 256 TLS1.2 Native AES SHA256 RSA 19: 53 AES256-SHA 256 TLS1 Native AES SHA RSA 20: 53 AES256-SHA 256 TLS1.1 Native AES SHA RSA 21: 53 AES256-SHA 256 TLS1.2 Native AES SHA RSA 22: 53 AES256-SHA 256 DTLS1 Native AES SHA RSA 23: 60 AES128-SHA256 128 TLS1.2 Native AES SHA256 RSA 24: 47 AES128-SHA 128 TLS1 Native AES SHA RSA 25: 47 AES128-SHA 128 TLS1.1 Native AES SHA RSA 26: 47 AES128-SHA 128 TLS1.2 Native AES SHA RSA 27: 47 AES128-SHA 128 DTLS1 Native AES SHA RSA 28: 10 DES-CBC3-SHA 168 TLS1 Native DES SHA RSA 29: 10 DES-CBC3-SHA 168 TLS1.1 Native DES SHA RSA 30: 10 DES-CBC3-SHA 168 TLS1.2 Native DES SHA RSA 31: 10 DES-CBC3-SHA 168 DTLS1 Native DES SHA RSA 32: 49200 ECDHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 ECDHE_RSA 33: 49199 ECDHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 ECDHE_RSA 34: 49192 ECDHE-RSA-AES256-SHA384 256 TLS1.2 Native AES SHA384 ECDHE_RSA 35: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1 Native AES SHA ECDHE_RSA 36: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1.1 Native AES SHA ECDHE_RSA 37: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1.2 Native AES SHA ECDHE_RSA 38: 49191 ECDHE-RSA-AES128-SHA256 128 TLS1.2 Native AES SHA256 ECDHE_RSA 39: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1 Native AES SHA ECDHE_RSA 40: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1.1 Native AES SHA ECDHE_RSA 41: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1.2 Native AES SHA ECDHE_RSA 42: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1 Native DES SHA ECDHE_RSA 43: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1.1 Native DES SHA ECDHE_RSA 44: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1.2 Native DES SHA ECDHE_RSA tmm --clientciphers 'DEFAULT:-3DES:!SSLv3:3DES+ECDHE:3DES+DHE:3DES+RSA' ID SUITE BITS PROT METHOD CIPHER MAC KEYX 0: 159 DHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 EDH/RSA 1: 158 DHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 EDH/RSA 2: 107 DHE-RSA-AES256-SHA256 256 TLS1.2 Native AES SHA256 EDH/RSA 3: 57 DHE-RSA-AES256-SHA 256 TLS1 Native AES SHA EDH/RSA 4: 57 DHE-RSA-AES256-SHA 256 TLS1.1 Native AES SHA EDH/RSA 5: 57 DHE-RSA-AES256-SHA 256 TLS1.2 Native AES SHA EDH/RSA 6: 57 DHE-RSA-AES256-SHA 256 DTLS1 Native AES SHA EDH/RSA 7: 103 DHE-RSA-AES128-SHA256 128 TLS1.2 Native AES SHA256 EDH/RSA 8: 51 DHE-RSA-AES128-SHA 128 TLS1 Native AES SHA EDH/RSA 9: 51 DHE-RSA-AES128-SHA 128 TLS1.1 Native AES SHA EDH/RSA 10: 51 DHE-RSA-AES128-SHA 128 TLS1.2 Native AES SHA EDH/RSA 11: 51 DHE-RSA-AES128-SHA 128 DTLS1 Native AES SHA EDH/RSA 12: 157 AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 RSA 13: 156 AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 RSA 14: 61 AES256-SHA256 256 TLS1.2 Native AES SHA256 RSA 15: 53 AES256-SHA 256 TLS1 Native AES SHA RSA 16: 53 AES256-SHA 256 TLS1.1 Native AES SHA RSA 17: 53 AES256-SHA 256 TLS1.2 Native AES SHA RSA 18: 53 AES256-SHA 256 DTLS1 Native AES SHA RSA 19: 60 AES128-SHA256 128 TLS1.2 Native AES SHA256 RSA 20: 47 AES128-SHA 128 TLS1 Native AES SHA RSA 21: 47 AES128-SHA 128 TLS1.1 Native AES SHA RSA 22: 47 AES128-SHA 128 TLS1.2 Native AES SHA RSA 23: 47 AES128-SHA 128 DTLS1 Native AES SHA RSA 24: 49200 ECDHE-RSA-AES256-GCM-SHA384 256 TLS1.2 Native AES-GCM SHA384 ECDHE_RSA 25: 49199 ECDHE-RSA-AES128-GCM-SHA256 128 TLS1.2 Native AES-GCM SHA256 ECDHE_RSA 26: 49192 ECDHE-RSA-AES256-SHA384 256 TLS1.2 Native AES SHA384 ECDHE_RSA 27: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1 Native AES SHA ECDHE_RSA 28: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1.1 Native AES SHA ECDHE_RSA 29: 49172 ECDHE-RSA-AES256-CBC-SHA 256 TLS1.2 Native AES SHA ECDHE_RSA 30: 49191 ECDHE-RSA-AES128-SHA256 128 TLS1.2 Native AES SHA256 ECDHE_RSA 31: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1 Native AES SHA ECDHE_RSA 32: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1.1 Native AES SHA ECDHE_RSA 33: 49171 ECDHE-RSA-AES128-CBC-SHA 128 TLS1.2 Native AES SHA ECDHE_RSA 34: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1 Native DES SHA ECDHE_RSA 35: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1.1 Native DES SHA ECDHE_RSA 36: 49170 ECDHE-RSA-DES-CBC3-SHA 168 TLS1.2 Native DES SHA ECDHE_RSA 37: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1 Native DES SHA EDH/RSA 38: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1.1 Native DES SHA EDH/RSA 39: 22 DHE-RSA-DES-CBC3-SHA 168 TLS1.2 Native DES SHA EDH/RSA 40: 22 DHE-RSA-DES-CBC3-SHA 168 DTLS1 Native DES SHA EDH/RSA 41: 10 DES-CBC3-SHA 168 TLS1 Native DES SHA RSA 42: 10 DES-CBC3-SHA 168 TLS1.1 Native DES SHA RSA 43: 10 DES-CBC3-SHA 168 TLS1.2 Native DES SHA RSA 44: 10 DES-CBC3-SHA 168 DTLS1 Native DES SHA RSA I added something else in there which I'll come back to later. Pitfalls As should be clear by now cipher configuration is a powerful tool, but as the song says, every tool is a weapon if you hold it right. And weapons are dangerous. With a little careless handling it is easy to lose a toe – or a leg. Whenever you are working with cipher suite configuration the old rule of 'measure twice, cut once' applies – and then double-check the work to be certain. There are several common pitfalls which await you. Misuse Perhaps the most common pitfall is simply misuse – using cipher suite configuration for that which it is not intended. And the single most common example of this comes from using cipher configuration to manipulate protocols. Given the keywords, as described above, it seems common for users to presume that if they want to disable a protocol, such as TLSv1.0, then the way to do that is to use a cipher suite keyword, such as !TLSv1. And, indeed, this may seem to work – but it isn't doing what is desired. The protocol is not disabled, only the ciphers that are supported for that protocol are. The protocol is configured on the VIP independently of the ciphers. !TLSv1 would disable all ciphers supported under the TLSv1.0 protocol, but not the protocol itself. Note that the protocol negotiation and the cipher negotiation in the SSL/TLS handshake are independent. What happens if the VIP only supports TLSv1.0/v1.1/v1.2 and the client only supports SSLv3 & TLSv1.0? Well, they'd agree on TLSv1.0 as the common protocol. The cipher list the client sends in the Client Hello is independent of the protocol that is eventually negotiated. Say the client sends AES128-SHA and the server has that in its list, so it is selected. OK, we've agreed on a protocol and a cipher suite – only the server won't do any ciphers on TLSv1.0 because of '!TLSv1' in the ciphers configuration, and the connection will fail. That may seem like splitting hairs, but it makes a difference. If a scanner is looking for protocols that are enabled, and not the full handshake, it may still flag a system which has been configured this way. The protocol is negotiated during the SSL/TLS handshake before the cipher is selected. This also means the system is doing more work, as the handshake continues further before failing, and the log messages may be misleading. Instead of logging a protocol incompatibility the logs will reflect the failure to find a viable cipher, which can be a red herring when it comes time to debug the configuration. The right way to do this is to actually disable the protocol, which doesn't involve the cipher suite configuration at all. For the control plane this is done through the ssl-protocol directive: tmsh list sys http ssl-protocol sys httpd { ssl-protocol "all -SSLv2 -SSLv3" } For example, if we wanted to disable TLSv1.0: tmsh modify sys http ssl-protocol 'all -SSLv2 -SSLv3 -TLSv1' tmsh list sys http ssl-protocol sys httpd { ssl-protocol "all -SSLv2 -SSLv3 -TLSv1" } For the data plane this can be done via the Options List in the SSL Profile GUI, via the No SSL, No TLSv1.1, etc. directives: Or via the command line: tmsh list ltm profile client-ssl Test options ltm profile client-ssl Test { options { dont-insert-empty-fragments } } tmsh modify ltm profile client-ssl Test options {dont-insert-empty-fragments no-tlsv1} tmsh list /ltm profile client-ssl Test options ltm profile client-ssl Test { options { dont-insert-empty-fragments no-tlsv1 } } The values are slightly different on the command line, use this command to see them all: tmsh modify ltm profile client-ssl <profile-name> options ? Use the right tool for the job and you'll be more likely to succeed. Truncation As I previously mentioned, in versions up to 11.5.0 the ciphers configuration string was truncated at 256 bytes. Starting in 11.5.0 that was increased to 768 bytes (see K11481: The SSL profile cipher lists have a 256 character limitation for more information), but that can still silently truncate long configurations. This is not a theoretical issue, we've seen users run into this in the real world. For example, little over a year ago I worked with a customer who was then using 11.4.1 HF8. They were trying to very precisely control which ciphers were enabled, and their order. In order to do this they'd decided to enumerate every individual cipher in their configuration – resulting in this cipher suite configuration string: TLSv1_2+ECDHE-RSA-AES256-CBC-SHA:TLSv1_1+ECDHE-RSA-AES256-CBC-SHA:TLSv1_2+ECDHE-RSA-AES128-CBC-SHA:TLSv1_1+ECDHE-RSA-AES128-CBC-SHA:TLSv1_2+DHE-RSA-AES256-SHA:TLSv1_1+DHE-RSA-AES256-SHA:TLSv1_2+DHE-RSA-AES128-SHA:TLSv1_1+DHE-RSA-AES128-SHA:TLSv1_2+AES256-SHA256:TLSv1_1+AES256-SHA:TLSv1_2+AES128-SHA256:TLSv1_1+AES128-SHA:TLSv1+ECDHE-RSA-AES256-CBC-SHA:TLSv1+ECDHE-RSA-AES128-CBC-SHA:TLSv1+DHE-RSA-AES256-SHA:TLSv1+DHE-RSA-AES128-SHA:TLSv1+AES256-SHA:TLSv1+AES128-SHA:TLSv1+DES-CBC3-SHA That string would save in the configuration and it was there if you looked at the bigip.conf file, but it was silently truncated when the configuration was loaded. Since this was 11.4.1, only the first 256 bytes were loaded successfully, which made the running configuration: TLSv1_2+ECDHE-RSA-AES256-CBC-SHA:TLSv1_1+ECDHE-RSA-AES256-CBC-SHA:TLSv1_2+ECDHE-RSA-AES128-CBC-SHA:TLSv1_1+ECDHE-RSA-AES128-CBC-SHA:TLSv1_2+DHE-RSA-AES256-SHA:TLSv1_1+DHE-RSA-AES256-SHA:TLSv1_2+DHE-RSA-AES128-SHA:TLSv1_1+DHE-RSA-AES128-SHA:TLSv1_2+AES256-S Note the last suite is truncated itself, which means it was invalid and therefore ignored. If their configuration had worked they would've had nineteen protocol+suite combinations – instead they had eight. Needless to say, this caused some problems. This customer was missing ciphers that they expected to have working. That is bad enough – but it could be worse. Let's imagine a customer who wants to specify several specific ciphers first, then generally enable a number of other TLSv1.2 & TLSv1.1 ciphers. And, of course, they are careful to disable dangerous ciphers! TLSv1_2+ECDHE-RSA-AES256-CBC-SHA:TLSv1_1+ECDHE-RSA-AES256-CBC-SHA:TLSv1_2+ECDHE-RSA-AES128-CBC-SHA:TLSv1_1+ECDHE-RSA-AES128-CBC-SHA:TLSv1_2+DHE-RSA-AES256-SHA:TLSv1_1+DHE-RSA-AES256-SHA:TLSv1_2+DHE-RSA-AES128-SHA:TLSv1_1+DHE-RSA-AES128-SHA:TLSv1_2:TLSv1_1:!RC4:!MD5:!ADH:!DES:!EXPORT OK, that looks fairly solid, right? What do you suppose the problem with this is? This is the problem; in 11.4.1 and earlier it would truncate to this: TLSv1_2+ECDHE-RSA-AES256-CBC-SHA:TLSv1_1+ECDHE-RSA-AES256-CBC-SHA:TLSv1_2+ECDHE-RSA-AES128-CBC-SHA:TLSv1_1+ECDHE-RSA-AES128-CBC-SHA:TLSv1_2+DHE-RSA-AES256-SHA:TLSv1_1+DHE-RSA-AES256-SHA:TLSv1_2+DHE-RSA-AES128-SHA:TLSv1_1+DHE-RSA-AES128-SHA:TLSv1_2:TLSv1_1: All of the exclusions were truncated off! Now we have the opposite problem – there are a number of ciphers enabled which the customer expects to be disabled! And they're BAD ciphers – ADH, DES, MD5, RC4. So this customer would be at high risk without realizing it. Be aware of this; it is very sneaky. The configuration will look fine; the truncation happens in the code when it loads the configuration. This is also one reason why I always recommend listing your exclusions first in the configuration string. Then you can never accidentally enable something. Unintended Consequences Let's say a new CVE is announced which exposes a very serious vulnerability in SSLv3 & TLSv1.0. There is no way to mitigate it, and the only solution is to limit connections to only TLSv1.1 & TLSv1.2. You want a cipher configuration to accomplish this. It seems straight-forward – just configure it to use only ciphers on TLSv1.1 & TLSv1.2: tmsh modify ltm profile client-ssl <profile> ciphers 'TLSv1_2:TLSv1_1' Congratulations, you've solved the problem. You are no longer vulnerable to this CVE. You know there is a but coming, right? What's wrong? Well, you just enabled all TLSv1.2 & TLSv1.1 ciphers. That includes such gems as RC4-MD5, RC4-SHA, DES, and a few ADH (Anonymous Diffie-Hellman) suites which have no authentication. As recently as 11.3.0 you'd even be enabling some 40-bit EXPORT ciphers. (We pulled them out of NATIVE in 11.4.0.) So you just leapt out of the frying pan and into the fire. Always, always, always check the configuration before using it. Running that through tmm --clientciphers 'TLSv1_2:TLSv1_1' would've raised red flags. Instead, this configuration would work without causing those problems: tmsh modify ltm profile client-ssl <profile> ciphers 'DEFAULT:!TLSv1:!SSLv3' Another option, and probably the better one, is to disable the SSLv3 and TLSv1.0 protocols on the VIP. As I discussed above. Of course, you can do both – belt and suspenders. And just to show you how easy it is to make such a mistake, F5 did this! In K13400: SSL 3.0/TLS 1.0 BEAST vulnerability CVE-2011-3389 and TLS protocol vulnerability CVE-2012-1870 we originally had the following in the mitigation section: Note: Alternatively, to configure an SSL profile to use only TLS 1.1-compatible, TLS 1.2-compatible, AES-GCM, or RC4-SHA ciphers using the tmsh utility, use the following syntax: tmsh create /ltm profile client-ssl <name> ciphers TLSv1_1:TLSv1_2:AES-GCM:RC4-SHA Yes, I had this fixed long ago. Remember back in the section on keywords I had this comparison example: tmm --clientciphers 'DEFAULT' tmm --clientciphers 'DEFAULT:-3DES:!SSLv3:3DES+ECDHE:3DES+DHE:3DES+RSA' Who caught the '!SSLv3' in the second line? Why do you think I added that? Did I need to? Hint: What do you think the side effect of blanket enabling all of those 3DES ciphers would be if I didn't explicitly disable SSLv3? Cipher Ordering In SSL/TLS there are two main models to the cipher suite negotiation – Server Cipher Preference or Client Cipher Preference. What does this mean? In SSL/TLS the client sends the list of cipher suites it is willing and able to support in the Client Hello. The server also has its list of cipher suites that it is willing and able to support. In Client Cipher Preference the server will select the first cipher on the client's list that is also in the server's list. Effectively this gives the client influence over which cipher is selected based on the order of the list it sends. In Server Cipher Preference the server will select the first server on its own list that is also on the client's list. So the server gives the order of its list precedence. BIG-IP always operates in Server Cipher Preference, so be very careful in how you order your cipher suites. Preferred suites should go at the top of the list. How you order your cipher suites will directly affect which ciphers are used. It doesn't matter if a stronger cipher is available if a weak cipher is matched first. HTTP/2 How is HTTP/2 a pitfall? The HTTP/2 RFC7540 includes a blacklist of ciphers that are valid in TLS, but should not be used in HTTP/2. This can cause a problem on a server where the TLS negotiation is decoupled from the ALPN exchange for the higher level protocol. The server might select a cipher which is on the blacklist, and then when the connection attempts to step up to HTTP/2 via ALPN the client may terminate the connection with extreme prejudice. It is well known enough to be called out in the RFC – Section 9.2.2. F5 added support for HTTP/2 in 12.0.0 – and we fell into this trap. Our DEFAULT ciphers list was ordered such that it was almost certain a blacklisted cipher would be selected.; This was fixed in 12.0.0 HF3 and 12.1.0, but serves as an example. On 12.0.0 FINAL through 12.0.0 HF2 a simple fix was to configure the ciphers to be 'ECDHE+AES-GCM:DEFAULT'. ECDHE+AES-GCM is guaranteed to be supported by any client compliant with RFC7540 (HTTP/2). Putting it first ensures it is selected before any blacklisted cipher. 3DES Back in the section on ciphers I mentioned that we label 3DES as being 168-bit, but that it only provides the equivalent of 112-bit strength. So, what did I mean by that? DES operates on 64-bit data blocks, using 56-bits of key. So it has a strength of 2 56 . 3DES, aka Triple DES, was a stop-gap designed to stretch the life of DES once 56-bits was too weak to be safe, until AES became available. 3DES use the exact same DES cipher, it just uses it three times – hence the name. So you might think 3x56-bits = 168-bits. 2 168 strong. Right? No, not really. The standard implementation of 3DES is known as EDE – for Encrypt, Decrypt, Encrypt. (For reasons we don't need to get into here.) You take the 64-bit data block, run it through DES once to encrypt it with K 1 , then run it through again to decrypt it using K 2 , then encrypt it once again using K 3 . Three keys, that's still 168-bits, right? Well, you'd think so. But the devil is in the (implementation) details. First of all there are three keying options for 3DES: - Keying option 1: K1, K2, K3 – 168 unique bits (but only 112-bit strength!) - Keying option 2: K1, K2, K1 – 112 unique bits (but only 80-bit strength!) - Keying option 3: K1, K1, K1 – 56 unique bits, 56-bit strength (Equivalent to DES due to EDE!) F5 uses keying option one, so we have 168-bits of unique key. However, 3DES with keying option one is subject to a meet-in-the-middle cryptographic attack which only has a cost of 2 112 . It has even been reduced as low as 2 108 , as described in this paper. So it does not provide the expected 168-bits of security, and is in fact weaker than AES128. To add some confusion, due to an old issue we used to describe 3DES as being 192-bit. See: K17296: The BIG-IP system incorrectly reports a 192-bit key length for cipher suites using 3DES (DES-CBC3) for more details. Of course, with the appearance of the Sweet32 attack last fall I would encourage everyone to disable 3DES completely whenever possible. We're also seeing a growing number of scanners and audit tools recategorizing 3DES as a 'Medium' strength cipher, down from 'High', and correspondingly lowering the grade for any site still supporting it. If you don't need it, turn it off. See K13167034: OpenSSL vulnerability CVE-2016-2183 for more information. Conclusion Believe it or not, that's the quick overview of cipher suite configuration on BIG-IP. There are many areas where we could dig in further and spend some time in the weeds, but I hope that this article helps at least one person understand cipher suite configuration better, and to avoid the pitfalls that commonly claim those who work with them. Additional Resources This article is by no means comprehensive, and for those interested I'd encourage additional reading: BIG-IP SSL Cipher History by David Holmes, here on DevCentral Cipher Rules And Groups in BIG-IP v13 by Chase Abbott, also on DevCentral OpenSSL Cipher Documentation K8802: Using SSL ciphers with BIG-IP Client SSL and Server SSL profiles K15194: Overview of the BIG-IP SSL/TLS cipher suite K13163: SSL ciphers supported on BIG-IP platforms (11.x - 12.x) K13156: SSL ciphers used in the default SSL profiles (11.x - 12.x) K17370: Configuring the cipher strength for SSL profiles (12.x) K13171: Configuring the cipher strength for SSL profiles (11.x) K14783: Overview of the Client SSL profile (11.x - 12.x) K14806: Overview of the Server SSL profile (11.x - 12.x)
