#infosec #cloud CloudPassage digs a bit deeper into the issue of security and public cloud computing and finds some interesting results
Security is a pretty big word. It’s used to represent everything from attack prevention to authentication and authorization to securing transport protocols. It’s used as an umbrella term for such a wide variety of concerns that it has become virtually meaningless when applied to technology.
For some time, purveyors of security studies have asked the market, “What’s stopping you from adopting cloud?” Invariably one of the most often cited show-stoppers is “security.” Pundits raced to tell us this, but in no wise did they offer deeper insight into what, exactly, security meant.
So it was nice to see CloudPassage dig deeper into “security in the cloud” with a recent survey it conducted. You may recall that CloudPassage has a more than passing interest in cloud-based security, as its focus is on cloud-based security with an emphasis on host-based firewalls. Published in February 2012, it sheds some light on what IT professionals consider most important with respect to public cloud security.
Not unsurprisingly, “lack of perimeter defenses and/or network control” was the most often cited concern with respect to security in public cloud environments with 25% of respondents indicating it was troubling. This response would appear to go hand in hand with the 12% who cited an inability to leverage “enterprise security tools” in public cloud environments. It is no secret that duplicating security architectures and processes in the cloud is not something we seen done at this juncture. When you combine an inability to replicate security policy and process in the cloud due to incompatibilities of infrastructure and software with a less than robust security service offering in public cloud environments, the “lack of perimeter defenses and/or network control” answer being top of the list makes a lot of sense.
There are myriad surveys that indicate organizations are moving to use public cloud computing, despite these concerns, and one assumes that this means they are finding ways to resolve these issues. Many organizations are turning back the clock and taking advantage of agent-based (host deployed) solutions to secure their assets in public cloud environments, which affords much better protection than nothing at all, and others still are leveraging the tried-and-true “checklist” method: manually securing servers based on best-practices and corporate policy.
Neither is optimal from an operational perspective. Neither is the use of cloud provider offered services such as Amazon security groups because the result is a disjointed set of security policies across multiple environments. Policy languages and implementation – not to mention capabilities – vary widely from service to service. While the most basic of protections – firewalling – is more compatible from the perspective of ability to codify, still the actual policy language will differ. These disconnects can lead to gaps in security policies that leave open to attack the organization’s assets. Inconsistent management and deployment processes spanning multiple environments leave open the possibility of human error and misconfiguration, an often cited cause of outages and breaches in general.
Where we are today is sitting with a disjointed set of options from which to choose, and the need to somehow cobble together these disparate tools and services into a comprehensive security strategy capable of consistently securing servers, applications, and other resources from attack, exploitation, and breach.
It is not really an inspiring view at the moment.
Vendors and providers need to work toward some common language and services that enable consistent replication – and thus enforcement - of the policies that govern access and protection of all corporate resources, regardless of location. Whether through standards initiatives or brokerage of APIs or better ability of organizations to deploy security solutions in both the data center and public cloud environments is not necessarily the question. The question is how can enterprises better address the specific security-related concerns they have regarding public cloud deployments in a way that minimizes risk of misconfiguration or gaps in policy enforcement while providing for operationally consistent processes that ensure the benefits of public cloud computing are not lost.
One of the interesting trends that we’re seeing is around the demand for consistency in infrastructure across environments, and this will eventually drive demand for integration of what are today “cloud only” solutions back into data center components. Folks like CloudPassage and other cloud-focused systems that deliver host-based security coupled with a SaaS management model will eventually need to consider integration with “traditional” enterprise solutions as a means to deliver the consistency necessary to maintain cloud-related operational benefits.
Right now we’re seeing a move toward preserving operational consistency through replication of policy from within the data center out, to the cloud. But as cloud-hosted solutions continue to mature and evolve, one would expect to see the ability to replicate policy in the other direction – from the cloud back into the data center. This is no trivial task, as it requires the SaaS management component of such solutions to become what might be considered a policy broker; that is, their system becomes the point of policy creation and management and it is through integration with both cloud and data center infrastructure that such policies are deployed, updated, and managed.
This is why the notion of API-enabled infrastructure, a.k.a. Infrastructure 2.0, is so important. It’s not just about creating a vibrant and healthy ecosystem of solutions within the data center, but in the cloud and in between, as well. It is the glue that will integrate disparate systems and normalize policies across environments, and ultimately provide the market with a broader set of choices that can more efficiently and effectively address the specific security (and other operational) concerns that may be preventing organizations from fully embracing cloud computing.