The National Hockey League's Tampa Bay Lightning added a Pillar Data Systems SAN disk array and InMage Systems...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Inc. disaster recovery software as key pieces of a data center overhaul aimed at offering better support for its ticket sales operation.
Ian Steele, director of IT for the Tampa Bay Lightning, said the direct-attached storage (DAS) setup he inherited last year wasn't sufficient to support key business processes.
"What drives our business is selling tickets," he said. "In order to sell tickets, our staff needs quick access to information. We store our CRM, but also data related to the building, statistics, several databases that power our sales force and, of course, email. [Microsoft] Exchange is our biggest application here with a couple of SQL applications running close second. We also have some video and bunch of graphics on our marketing side, and store all of that on a SAN."
Steele said he decided to consolidate the team's infrastructure to four physical servers with VMware, and that required networked storage. "The first step in any data center consolidation is nailing down your storage," he said. "We wanted to go virtual, but we needed a solid storage foundation to build a VMware environment on."
He said he looked at the large storage vendors, including EMC Corp., Hewlett-Packard (HP) Co. and NetApp. "None left a good feeling in my mouth," he said. "They were just after the initial sale and they weren't going to be with me as a partner through the years."
He said Pillar Data Systems took more of a partner approach. "I told them where I wanted to go from a virtualization standpoint," he said. "Not only did the product and technology seem to work, but Pillar seemed like a good company with its customer service, support and reliability. So far I've been very impressed."
He has also been impressed with Pillar's quality of service and thin provisioning features since buying a Pillar Axiom 500 SAN with 20 TB six months ago. Steele said he started with iSCSI for budget purposes, and is upgrading to Fibre Channel. "I wanted to do everything in steps because I can't get budget for everything at once," he said. "We hooked it up via iSCSI to our existing server and migrated all the data to the SAN. The Exchange server and SQL servers run off the SAN, with a couple of file servers serving files from the SAN."
His next step was to upgrade the network and virtual servers. For that, the team turned to Cisco Systems Inc.'s Unified Computing System (UCS) rackmount servers and VMware. "I didn't even know Cisco made servers," Steele said. "I had quotes from HP and we were ready to rip our stuff out and put in new machines. Cisco said 'We sell servers.' I spent a couple of months researching them. Their platform is built around VMware and designed to work off SAN storage. "
Steele is using VMware vSphere 4.0 with plans to upgrade to 4.1, and will upgrade to Windows Server 2008 and Exchange 2010.
He also bought two Cisco MDS 9124 Fibre Channel switches with eight 4 Gbps ports apiece. "We still have iSCSi going to the older servers and Fibre Channel to the new VMware environment," Steele said. "We'll keep iSCSI as we migrate data, but we made an investment in a Fibre Channel SAN and HBAs on our servers. There's no reason to continue to use iSCSI [long term]. "
The last piece of the overhaul will be disaster recovery. The team is searching for an out-of-town disaster recovery site and will replicate from its data center in the St. Pete Times Forum arena. "In the past, our DR didn't exist," Steele said. "We were backing up to tape and rebuilding servers if one went down." It was an hour-long restoration from tape if a server went down, he said.
He purchased InMage DR-Scout to protect virtual machines (VMs), Exchange and SQL. "I like that it's appliance based and doesn't sit in the path of the storage," Steele said. "It backs up VMs to a local repository, and replicates critical VMs, Exchange, SQL and file servers to an offsite DR site. We can fail over to that cold site if and when we have a disaster."
Steele hopes to have the DR site running by the end of August. "Our season starts Oct. 9 and we can't do anything once the season starts," he said.
Once the disaster recovery plans are set, Steele's major challenge will be keeping up with data growth. He said he has approximately 2 TB of data now, but his 120 users are rapidly expanding that footprint "They're asking me all the time for more space on the SAN," he said. "I told them to wait for the new environment. Now we're offloading video and images offline to keep off our network. We're looking at bringing on document images for business, and that will take up a tremendous amount of space.