Name: SCOTT ENGEL Director of IT infrastructure Transplace, a third-party logistics provider Dallas
Title:
Company:
Location:
Favorite data center product: Silver Peak Systems' NX-7600
We're doing near-real-time replication between our two data centers, in Dallas and Lowell, Ark., and we're seeing anywhere from three to five times reduction – and sometimes as high as 28 times -- on that traffic volume using the Silver Peak product. We have a single DS-3 today and we're able to push more than 1 terabyte of data through it daily.
Without the Silver Peak, we were going to have to upgrade to OC-3 from the DS-3. And, the ROI with the Silver Peak was only approximately one year. We've been using the Silver Peak for a year and a half, and we're replicating even more all the time, so now we're moving up to OC-3. But, again, if we didn't have the Silver Peak we would need to go to something even larger than that. We continue to add replication jobs to give us better disaster-recovery capabilities. We want as near real time as possible.
On deck: Cisco Nexus 1000V Series Switch
From a network side, we're looking at implementing the Cisco virtual switch, the 1000V, in the next couple of months. Our main driver is visibility.
Since our VMware roll out, our systems or virtual machine administrators have had control of the virtual switch within VMware. The network team's visibility into the virtual switch was limited, as was the ability to enforce security at that Layer 2 level. So we're moving to the Cisco 1000V so the networking group can manage and support that product with the VMware team and so we can get better visibility and security on those ports.
Today, if the VM hosts are talking from the same ESX server, that traffic isn't visible on the network – it's never put on the wire. But with the Cisco virtual switch, we'll have visibility into that traffic for troubleshooting and security purposes. So we'll be replacing the VMware virtual switches, dedicating a couple of ESX hosts to the 1000Vs. We'll have a redundant pair per VMware VirtualCenter [now vCenter Server]; today we have three VirtualCenters, so we'll have a total of six 1000Vs -- three primary, three passive/secondary.
Dream product: Real-time intelligent monitoring for small and midsize businesses
There used to be a network management company called SMARTS that offered an intelligent monitoring solution that would identify and allow drill down into single points of failure in real time. But the SMARTS product (acquired by EMC) was not affordable for the small-to-midsize business. So it would be really cool to have a monitoring/management tool that went across all portions of the infrastructure and allowed you to have a real-time, single point-of-failure solution. If I'm having a problem with a switch, it'll show me, 'Hey, you're getting whatever packet drops on this switch port,' or it'll show me where problems are with server CPUs or database configurations. There are several tools out there that are larger, but I don't see one that's scaled down for the small- and midsize business.