"Overloaded Networks and the Evolution of Rackonomics"
By Clabby Analytics

The existing model for data center network scaling is broken. New technologies (such as multi-core/multi-threading CPU architectures), new practices (such as server/storage consolidation and virtualization), and new application designs (such as input/output [I/O] intensive Web 2.0 applications and message-heavy Web services applications) have all contributed to a monstrously huge increase in demand for network bandwidth.
 

And to deal with this demand, IT (information technology) managers are being forced to continually buy and deploy super-expensive, power-hungry, high speed core switches.
 

This is no way to scale network bandwidth — it necessitates and perpetuates the continual deployment of costly, distributed core switches. Clearly, there’s got to be a better way…
 

BLADE Network Technologies (BLADE), a market-leading supplier of Gigabit and 10G Ethernet network infrastructure solutions, believes that it has found a better, more economical way to scale network bandwidth for consolidated rack/blade server/storage environments.
 

Rather than tying racks and blades into expensive, external core switches, BLADE has architected two new switch modules that can easily fit into existing blade server/rack environments — enabling networking bandwidth to be scaled cost effectively within a blade or rack cabinet, rather than externally.
 

Using this approach, data center architects can standardize on a unified and affordable rack-level network infrastructure to cost-effectively provision and scale out Web 2.0 environments, high-performance computing (HPC) clusters and virtualized data centers.
 

BLADE refers to this rack-based, economically differentiated approach as “Rackonomics”. In this Product Brief, Clabby Analytics examines BLADE’s “Rackonomics” concept (the idea that putting switches into blades and racks is more economical than using external core switches) — and its two new RackSwitch offerings (the G8000 and G8100) designed to implement its Rackonomics view.
 

What I find is that BLADE’s put-it-in-the-rack strategy is valid and well conceived — and will enable IT managers to save tens-of-thousands of dollars in switch hardware, deployment, and energy costs for every external core switch they are able to avoid deploying. Further, I find BLADE’s new RackSwitch designs to be highly innovative (especially when it comes to air flow design and energy efficiency); highly integrated (especially when it comes to the integration of disparate network fabrics); highly-virtualizable (if there is such a word); and extremely cost effective (especially when compared to CISCO core switching options).
 

View more presentations from bladenet.