Let's first get the definition of what a microburst is and why they can occur. A microburst occurs when a huge amount of data hits a network over a short period of time—like during time-sensitive market data updates.
Microbursts should not be disguised with lipstick by adding more bandwidth. You won't know what is actually behind what you initially see. Microbursts saturate the available bandwidth for your critical links and add a significant amount of latency to critical business transactions that need to occur without delay.
By hiding behind typical measurements for average throughput and utilization—microbursts cannot be detected over most one, five, or fifteen-second intervals. For example, any industry that relies on the timing of a transaction to affect business success, is affected by microbursts. When a desired transaction is caught up in a queue because of microburst saturation, you [or your automated trading program] don't make the trade, the sale, or whatever it is you need to execute during the time of saturation.
Key Concept—When you look at your link utilization or average throughput, you cannot see the problem but know that something is wrong—a transaction is not executed based on what you expected from your market data feed. The transaction could slip, not get executed, or be retransmitted and executed at a less optimal time.
You might consider allocating a bunch of extra bandwidth to the links affected and everything would run smoothly. But just how much? What if your microbursts were well over 1 Gb of data over a 1 Gb interface, and you didn't even know it? The only way you can detect microbursts is over milli-second intervals—duh!
Enter Cascade Shark. You can do a deep dive packet analysis on a link that you know runs the critical data but only shows typical measurements of say, just over 400 Mbits/sec of throughput. When you zero in on the time frame of interest to the actual packets, you see the microburst activity as shown here (saturating the link capacity at 1.00G Bits/s):
Riverbed OPNET AppResponse Xpert has been conveniently integrated with the deep dive packet analytics of Cascade Shark as on optional module in the latest RPM release. For more information, read this solution brief on the Cascade Shark module for AppResponse Xpert.
With Riverbed OPNET AppResponse Xpert and Cascade Shark, you don't have to do the guess work and you can make sure that the transactions you want to occur down to the millisecond timing of a trade—or whatever it is critical to your business—will occur. You know what we mean, we want you to have really really really really fast apps.
Thank goodness you don't have to put lipstick on a pig—or a microburst!
Riverbed Performance Management (Cascade & OPNET)
Riverbed Performance Management solutions provide IT with the visibility and actionable insight to help deliver the application performance that users and business demand.
#riverbed #casacade #shark #opnet #appresponse
Follow me on Twitter @Not2LateCapHill
Also, check out my previous post about setting your apps on fire.
Riverbed, at more than $1 billion in annual revenue, is the leader in application performance infrastructure, delivering the most complete platform for the hybrid enterprise to ensure applications perform as expected, data is always available when needed, and performance issues can be proactively detected and resolved before impacting business performance. Riverbed enables hybrid enterprises to transform application performance into a competitive advantage by maximizing employee productivity and leveraging IT to create new forms of operational agility. Riverbed’s 27,000+ customers include 97% of the Fortune 100 and 98% of the Forbes Global 100.