Scaling networks for internet traffic growth: application acceleration

Lee Chen, A10 Networks founder and CEO, discusses how financial institutions can leverage application acceleration technology to keep pace amid increasing Internet growth – today and in the future.

FST: With your company’s deep networking expertise, can you share the state of enterprise networking environments today?

Lee Chen: Over the past decade, we have seen rapid Web application traffic growth with Internet driven applications and media rich communications. Customers in all verticals are considering deploying or upgrading application acceleration technology to handle high volume network applications.

Networking and security vendors continue to develop application acceleration platforms with the performance needed to meet this growth so that applications are available and scalable with customers’ needs, today and in the future. With countless advances in networking technologies since the 1990s, one thing that has not changed is the ever increasing need for network and application speed.

FST: Can you share some perspective with the finance industry on application acceleration technology?

Lee Chen: If you are currently evaluating application acceleration solutions, you should consider three must-have requirements: Acceleration, Optimization and Security.

Application acceleration solutions should have the architecture to accelerate and improve business application response times. They should have the intelligence to optimize and reduce bandwidth, space and energy requirements. And they should have the high-performance SSL acceleration and other security features to address increasing security requirements for Web-based financial transactions.

There are other considerations such as reliability, availability, ease-of-use and ease-of-migration to lower operating expenses.

FST: So how does all this play out with today’s application acceleration solutions?

Lee Chen: Many of the incumbent vendors’ solutions are based on legacy architecture that is not designed to address the greater scale of Internet traffic and security requirements – today and tomorrow. In order to address the need for speed, vendors continue to move their single-CPU software solutions to more powerful multi-CPU application acceleration systems. But without the complete redesign of the software, performance benefits as a result of more powerful hardware will never be realized.

FST: Can you give an example of this current phenomenon in application acceleration?

Lee Chen: If you look at my former company Foundry Networks, Foundry stopped proprietary ASIC development within its systems several years ago. The industry has shifted focus from developing proprietary hardware to merchandized Layer 2 to Layer 3 and SSL acceleration hardware.

The battle now is to develop smart multi-threaded software to take advantage of the powerful multi-CPU hardware. Believe me, it is a real challenging task to develop super efficient multi-threaded software. Luckily for A10, we have a team of engineers with high-performance Layer 2-3 and Layer 4-7 networking, and super-computing system design experience.

FST: So what should the next generation application acceleration system include?

Lee Chen: The next generation application acceleration system should include a multi-CPU architecture. A properly designed multi-CPU system will have maximum per-CPU independence with almost no interference among CPUs. This is because a set of instructions running across an optimized multi-CPU system architecture is exponentially faster than the same set of instructions going across a single-CPU system.

Multiply these sorts of benefits at every network gateway, and you begin to see why we are so excited about this new generation of application acceleration switches that include multi-CPU architectures. With this next generation architecture, the nature of systems administration changes to where your administrator is solving traffic bottlenecks before they occur.

FST: Application acceleration is a very popular technology in the financial sector today. Can you share any advice on how to pick technology that can safeguard applications today and in the future?

Lee Chen: The biggest piece of advice is to work with competent technology practitioners. IT administrators need to choose technologies that can scale with their constantly and rapidly changing business requirements. So keep an eye on new innovative and flexible technologies to meet your business needs.

Also, be careful in choosing complex solutions to simple problems – and this includes solutions that are getting most of their Layer 4-7 application acceleration functions from complicated, hard-wired ASIC technologies. Getting simple feature requests into the accelerating ASIC requires 18 months or more of development time. IT administrators should think twice before picking such complex solutions.

Beyond that, pick technology vendors that will work with you to come up with the solutions for your business needs today and tomorrow. Who is to say what people are doing 5 years from now; 10 years from now? But one thing is for sure, in addition to the “need for speed”- Networks will continue to become bigger, more bandwidth intensive, more complex - and require sophisticated management tools to perform efficiently.