on 18-Jun-2013 12:07
Before June 29th, 2007, if you asked someone what kind of phone they had they would likely have answered with a manufacturer's name or brand: Motorola, Blackberry, Nokia. In September 2008, Google unleashed Android on the world and everything changed. Today, phones are primarily identified by platform - not manufacturer. It's Android versus Apple (mostly because iOS doesn't flow nearly as nicely from the tongue), a battle of technology titans with a market potential that is virtually unknowable because that market is determined by what the platform supports. At the moment, that's pretty much everything in the app world.
Regardless of which side of the fence you fall in the battle, it's hard to disagree that both are very successful, primarily because of the strength inherent in the platform approach. Platforms, it turns out, are a key success factor whether we're talking about phones or cloud or application delivery controllers (ADC).
F5's offering has been, since 2004, a platform supporting an extensible ecosystem of application-focused "apps" (we call them modules; browsers call them plug-ins). That platform is TMOS and, much like iOS, spent a lot of its life tied to proprietary, purpose-built hardware. F5 is committed to TMOS on existing and forthcoming new hardware form factors, but the adoption of cloud and virtualization and ongoing transformations in the operational realm of application deployment such as that of DevOps, made it clear that other form factors were necessary. Thus, F5 made TMOS and its ecosystem of application delivery modules available in a virtual form factor across all hypervisors, a unique approach in the ADC market. As demand for architectural and feature parity across the cloud grew, the market responded nearly in concert by enabling respective offerings for deployment within AWS. Despite claims to the contrary, competing offerings were announced at the same time at Amazon's inaugural conference, re: Invent in 2012. The "virtual" ADC market (if there even is such a thing separate from the ADC market) is just emerging, with a variety of different needs and requirements across audiences. No doubt its infancy and difficulty in tracking has been a factor in the diversity of forecasts with respect to the size and growth of the market.
F5 customers have consistently required platform parity of features whether deployed in the cloud or in the data center. To offer a less capable platform as many competitors have done would prevent the market from realizing the full value of the platform to make applications secure, fast, and available no matter where they are deployed.
Over the ensuing years, however, there's been a significant divergence of approach to application delivery across the market. F5 remains committed to a robust platform deployable in a variety of form factors supporting a diverse ecosystem of application delivery capabilities; capabilities that have continued to evolve as the needs of applications and architectures of data centers continue to change. A comparison of the initial definition of ADC (then called “Web Optimization”) to its modern incarnation shows clearly that while some capabilities mature, other capabilities continue to be added or evolve along with the applications they are critical to delivering. The same comparison also shows that F5 introduces these features and functions on average 5 years before they become part of the industry accepted ADC definition.
Rapid support for network and application protocols such as SAML 2.0, VXLAN, NVGRE, and SPDY across all ADC capabilities is a good example of how a platform provides a robust and evolving base upon which application delivery capabilities, like extending the ability of BIG-IP to manage access for an always-in-flux mobile device market, can be deployed. Eventually, such ecosystem-delivered capabilities become a part of the ADC definition - a trend we're starting to see today with WAN optimization and anticipate seeing with web application acceleration capabilities and emerging SDN-related protocols.
There is no doubt that capabilities not associated with ADC today will eventually be assimilated into ADC and expand its definition again. Consider the impact of APIs - a multi-billion dollar market today - on application delivery needs. The capabilities required to meet such needs will naturally gravitate toward the most strategic point of control in the data center - the ADC - in order to facilitate emerging challenges with version management, metering, performance, and security. F5 ADC is a platform, not a product, and as such its capabilities are not fixed. It easily adapts to whatever changes may occur within the data center and beyond. This is a core reason why the ADC market continues to expand over time; as some functions become commoditized new features and capabilities are added based on demands and application trends. There isn't a finite pool of ADC customers out there because F5 ADC keeps changing, thus broadening its applicability to new markets.
Over the past 10 years, the ADC market has continued to experience growth. Every time a new capability is integrated into the ADC, such as WAN optimization or web application firewalling or application acceleration, the market expands. The upper limit on the ADC market is constrained only by current definitions; definitions that historically evolve and expand.
ADC must be a platform and not a product to continue to adapt to the needs of applications and businesses over time. We've heard a lot about SDN (software-defined networking) and its ability to reduce time-to-market for new protocols and network functions by treating the SDN controller as a platform, extensible through API-integrated "applications". The concept of being a platform and promoting an ecosystem approach is critical for future success of SDN. That concept has been embodied in TMOS since 2004 and the platform-ecosystem approach is key to ensuring that our products are able to continue to evolve and respond rapidly to market and architectural needs by extending the ecosystem to include whatever new capability might be necessary.
While some are enamored of the ability to physically consolidate application delivery services on a single device, many more appreciate the operational consistency that comes from deploying a variety of what seem to be very different capabilities - access control, network and application security, SSL termination, application acceleration, VDI*, load balancing and DNS - on the same programmable platform because it means scale and management are consistent across the ecosystem.
TMOS is a platform on which a significant selection of application and data center needs are met and will continue to be met precisely because of F5's commitment to making it available on a variety of form factors (hardware, virtual, and software) along with a robust, adaptable ecosystem. When a new application delivery challenge arises, the F5 ADC ecosystem will incorporate new tools to meet that challenge even when it does not fall into the existing definition of an application delivery controller.
ADC is a platform, and as such it is not constrained by definitions and checklists of capabilities. Its ecosystem of capabilities will continue to evolve and transform as needed by the applications it delivers.
* Yes, even Citrix. In fact, when a customer looks at both NetScaler and BIG-IP to support their Citrix XenApp environment, BIG-IP is chosen 71% of the time.