It’s hard enough trying to cope with an aging server and storage infrastructure that’s near its end of life, but it’s even worse having everything spread across five data centers.
For several years, that was the reality for the city of Atlanta’s information management team. “We were having all kinds of issues keeping production going,” says Raymond Brooks, director of information management infrastructure. “We had system outages every single day and a lot of application interruptions for our customers. It was really crazy.”
In 2013, Brooks and his team decided the time had finally come to modernize, virtualize and consolidate the city’s data center operations. But instead of going down the traditional path of researching, purchasing and integrating individual components, they chose a faster, easier and less expensive solution: a converged infrastructure. Converged infrastructure bundles server, storage, networking, virtualization and management into a single preconfigured platform.
Converged infrastructure simplifies decision-making, streamlines implementation and offers greater performance and scalability than a custom solution, says David Cappuccio, distinguished analyst for data centers with Gartner. The solutions also provide a stepping stone for offering cloud computing services.
“With converged infrastructure, IT teams don’t have to worry as much about what the components are and how they work together — their real concern is: Do I trust the vendor that’s supporting everything, and will they be there for me?” Cappuccio explains.
Converged Infrastructure Is More Versatile
Atlanta selected EMC’s VSpex modular solution that features Cisco Nexus switches and Unified Computing System server blades, EMC VSpex storage and VMware vSphere virtualization and management.
Though Brooks expects the new solution to deliver numerous benefits, he and his team initially pursued converged infrastructure for its single-source support.
“We get everything we need built into one,” Brooks says. “So if we have any issues, instead of having to call three or four vendors to try and troubleshoot the problem, now we are able to call EMC and receive full support for everything. That was very appealing.”
Atlanta rolled out its first VSpex units 18 months ago as a first step in its effort to modernize and consolidate five data centers into two — one for primary production and one to serve as a backup and failover site.
Though implementation presented a learning curve, the rollout was seamless. “It was very simple, and we were able to get it implemented within budget and in the time frame we needed,” Brooks says.
Atlanta’s IT team and its customers appreciate the speedy deployment afforded by the converged infrastructure. “We can roll out applications or turn around requests much, much quicker than we used to,” Brooks says. “We can spin up a server in a matter of minutes.”
Moving from tape backup to data deduplication also speeds backup and recovery. “We’re able to turn around our backup and restore in minutes, as opposed to days. Instead of taking 24 to 36 hours to perform all our backups, we can get it done in three or four hours,” Brooks says.
For the city of Miami Beach, Fla., Cisco’s Unified Computing System fulfilled key requirements for a new enterprise resource planning (ERP) application. Featuring Cisco blade- and rack-mounted servers and Nexus switches, the UCS solution gave the city the ability to run a mixed environment of virtual and physical machines, while also enabling the team to consolidate rack space and upgrade to a 10 Gigabit Ethernet network.
The first UCS units have been in place for only a few months, but the city’s IT support division has already realized cost and time savings.
“The window from unboxing to production has shrunk considerably for us,” says Jose Suarez, IT support division director for Miami Beach. “We can send a single administrator out to literally plug a server in and walk away.”
Frank Quintana, IT systems support manager for Miami Beach, says the upfront work done with converged infrastructure enables savings and performance benefits to increase exponentially as time goes on.
“The way the system is architected, it allows you to set up cookie cutter templates,” Quintana says. “For example, if you want a server to be a VMware host, you assign the VMware host template, and the system applies all the preconfigured elements that are part of that template. So it handles the networking, it attaches the storage, installs the operating system — it all happens automatically because of the work you did on the front end.”
The ERP rollout has gone so well that as the older virtual infrastructure ages, the IT team will move it to the UCS platform. Ultimately, Suarez anticipates being able to reduce the city’s server footprint by 75 percent and its energy consumption by half.
Early Advantages of Converged Infrastructure
Robust scalability and resiliency is another key benefit of a unified platform, which is why the state of Alaska decided to rely on Cisco’s FlexPod converged infrastructure as the foundation for a data center centralization initiative. FlexPod includes Cisco UCS blade servers and switches, NetApp storage products and VMware virtualization.
Until recently, Alaska’s agencies operated their own data centers, a strategy that has resulted in servers being spread among more than 180 rooms and closets throughout the state, says Jim Bates, Alaska’s director of enterprise technology services. Centralizing the data center enables the IT department to provide a variety of managed and unmanaged services. But because there’s no mandate for agencies to cede local control, Bates and his team have to really prove the value of centralization.
Already, there have been successes. When a third-party vendor failed to successfully implement the state’s Medicaid eligibility system, the Department of Health and Social Services asked Bates and his team to take over. “We stood up 22 FlexPod servers in our virtual environment in a week,” Bates says. “And we bailed the agency out, saving them about $140,000 a year.”
In June, the team completed the migration of Alaska’s ERP and accounts payable systems from legacy mainframes to the FlexPod platform.
“With those two projects alone, we’ve really been demonstrating the ability of the system to be able to scale and adapt very quickly,” Bates says, noting that the data center consolidation and modernization effort will be instrumental in satisfying the state legislature’s push to save $60 million in IT spending over the next three years. “That has started getting people’s attention and helping them realize that our data center and the FlexPod platform is not just a viable option, but really a better option for them.”