Network Managers: Know Thyself
Posted by
Robert Mullins
February 09, 2012
The transition to 10 Gbit Ethernet (GbE) from 1 GbE appears to be gaining traction, 4G and LTE wireless networks are quickly replacing 3G, and the introduction of Intel's Romley microprocessor platform later this year, with built-in 10-GbE support, is expected to drive another round of server refreshes industrywide. All of this is causing network operators to buckle up for a bumpy ride as they try to manage the torrent of data on their networks these innovations will unleash.
And even as 10GbE is advancing, an increase to 40GbE and 100GbE connectivity is generating the next threatening waves. But many of the clients that industry researcher Zeus Kerravala, principal analyst at ZK Research, consults for don't have a clear sense of what traffic courses through their networks and how to best manage it.
"One piece of advice I always give network managers is don't re-architect your network or make any decisions until you truly know your network," he says. "My estimate is that only about 25% of network managers really know their network."
Enter Gigamon, which just introduced an additional network chassis that delivers what is calls a Traffic Visibility Fabric technology. It sits on top of the network layer as an abstracted layer, providing that visibility for various network management tools that track performance, quality of service, security and other important metrics.
The GigaVUE-HD4 Traffic Visibility Node chassis is a five-rack unit (5RU) device that complements the GigaVUE-HD8 chassis unveiled by Gigamon in May 2011. That is a larger 14RU chassis that customers thought was great, but a bit much.
"We designed this [HD4] box because customers were saying your big chassis, the HD8, is pretty efficient on space at 14RU, but it's big. Is there something that fits between a classic 1RU and the 14RU?" says Paul Hooper, network visibility strategist for Gigamon.
The HD4 offers eight 40-Gbps ports, 96 10-Gbps ports or 176 1-Gbps ports, depending on how the blade is configured. It delivers up to 1.3 Tbits of throughput and aggregates data from a row of servers by sitting at the end of each row.
Page: 1 | 2 | Next Page »
