Brocade wants to give Fibre Channel storage infrastructure analytics and monitoring of traffic between servers and storage, to help benchmark application performance and diagnose application problems.
As Brocade's Senior Director for Australia and New Zealand Gary Denman explained to Vulture South, it's particularly useful in outsourced and hosted environments, when both the customer and provider need to match application performance against SLAs.
The Brocade Analysis Monitoring Platform is a 24-port, 2U piece of iron that includes two CPUs for frame processing, running the company's Fabric OS as the basis for the analytics and integration with Brocade Network Advisor.
Phil Coates, systems engineer manager, A/NZ, said the aim is to build on the company's existing Fabric Vision platform with stateful analysis that can analyse up to 20,000 flows and millions of IOPS simultaneously.
“Each flow is an application, or people connecting to that application,” Coates explained.
To get around privacy concerns about traditional network “taps” for flow monitoring, Coates emphasised that “we're not capturing the data.”
“We're looking at how the application is performing for a client,” he said.
There's another reason to avoid “taps”, he added: if a new server is spun up, there's a lot of work to add the server to what's monitored.
The Brocade Analytics Monitoring Platform provides a “non-intrusive connection” that can make a virtual connection to “any port in the fabric” he said.
“I can now look at a frame going into a server, see the response coming out of that server, and tell how long it's taken to process.”
By capturing the latency of an application across server, fabric and storage, the user also gets historical analysis for capacity planning purposes.
Integration into the VMWare platform helps “engage with the application layer”, Coates added.
Brocade's announcement is here. ®