When we look at where the networking industry currently is, one could easily use words like angst and uncertain to describe where traditional networking architectures have been left after server virtualization.
Where in the past there were 100 physical servers for 100 applications, in today’s data center the same 100 applications may reside on 6-10 physical servers, and in many cases these servers are in a blade form factor in which many can share the same chassis.
What we are seeing is that the natural flow of data and packets has gone from a north-south relationship to one that is of an east-west nature (if not encumbered by traditional network architectures). This has put massive strain on existing networks and forced clients and vendors alike to try and come up with creative ways to solve this growing problem before it threatens the service levels and availabilities that networks need to provide.
A new technology called IRF, or the Intelligent Resilient Framework is rising to solve this challenge. IRF allows users to group switches together to be managed as one large, virtualized single switch. It also eliminated the requirement of using Spanning Tree Protocol (STP) to ensure availability. STP is a time consuming planning task, and also renders upwards of 50% of the links to nothing more than a passive failover role. So in the end clients are spending precious time on provisioning networks that only achieve half of their potential. When you couple this with the increased density in required bandwidth that virtual machines require, and the fact that most companies still have a north-south traffic flow, we start to see that networks in their current form are not adequate for the changes that have hit them over the past 3-4 years.
IRF is a 100% FREE feature and all it requires is a standard 10GbE cable to connect one switch to another, and now both switches can be managed as one, pushing data through 100% of ALL links, rather than just the active links as seen with STP in traditional environments. This means organizations can spends up to 50% less time provisioning their networks, increase resiliency, and double performance.