City of Terre Haute Case Study

Fast Facts

The City of Terre Haute’s IT department supports over 600 distributed city government, emergency, transportation and utility staff that serve a growing population. When a major flood exposed the need for a disaster recovery plan, IT was tasked with renovating their infrastructure to ensure it could survive a disaster. To achieve this, they needed a new storage system that would fit the taxpayer’s budget, provide replication to a remote site and achieve a recovery time objective of 6-8 hours.

In a comparison of five vendors, only Scale Computing was able to meet all of the city’s requirements within their budget.

Introduction

Terre Haute, Indiana is the largest city in Indiana’s Wabash valley and the 12th largest city in the state. As home to five major universities, an international airport and several major manufacturing facilities, the city has seen a steady 2 percent increase in population over the last five years. The City of Terre Haute’s Information Technology department supports the growing population with over 600 remote IT users that include Fire and Police departments, Parks, Streets, Transit, Housing, Utilities and city government.

Challenge

When the Wabash river swelled to record levels in 2008 and flooded portions of the city, it exposed a weakness in the city’s public safety program. Emergency services that relied on IT data and applications to marshal and dispatch resources were threatened and the IT department realized it had no disaster recovery plan and no safe place for critical servers.

When Police and Fire get dispatched, they need as much data as possible in the field. Our police force has 85 cars with PCs and they need the ability to access state and federal databases, especially in an emergency.

―Bradley Speidel, Director of Information Technology, City of Terre Haute

Near the end of 2009, Bradley Speidel, Director of Information Technology, began to plan a new infrastructure, virtualizing servers and implementing desktop server virtualization for the police department. Speidel and his team migrated data and applications from HP G3 physical servers, which had been in service since 2005, to VMware virtual machines. Speidel then hit a roadblock with his existing HP MSA storage. The original HP MSA1000 had been in production for over five years and had limitations on port speed, port density and could only use Fibre Channel for connectivity. His second MSA 1000, originally purchased for lab/development use and forced into production, had even fewer ports, was nearing capacity, and did not offer an integrated ability to replicate data to a secondary site. Speidel started the search for a replacement storage system that could protect the city’s critical data and applications and could grow economically, as the city’s storage needs increased – all for $120,000.

Solution

With the intent of getting the most out of tax payer dollars without cutting corners on storage performance, capacity and support, Speidel put out an RFP that included requirements for 15 terabytes of storage at two different sites, with the ability to replicate between them. Speidel also wanted to invest in five years of maintenance up front, as a hedge against shrinking city budgets. Five vendors responded to the RFP. Of the respondents, Dell’s Compellent Storage Center was quickly eliminated based on price. Equallogic storage from Dell was also eliminated on price after the inclusion of five years of support. NetApp proposed a solution closer to Speidel’s budget, but would require a secondary system less capable than the primary site. HP found a way to meet the budget and extended support requirements with their P4000 SAN solution, formerly known as Lefthand, but heavily pushed a synchronous replication solution between sites that Speidel’s team didn’t really need and had reservations about managing. The Lefthand solution also only provided 12 terabytes of storage per site with a substantial cost to add more capacity later. Frank Leonard of Leonard-McDowell proposed a Scale Computing storage solution consisting of two storage clusters with built-in asynchronous replication. Scale’s storage portfolio is based on the company’s Intelligent Clustered Operating System Storage (ICOS) technology, which enables users to build a storage system out of multiple storage nodes, and grow the system as needed - without suspending services or migrating data. IT managers are able to build out storage clusters, starting with just six terabytes (raw). For redundancy, data is striped and mirrored across multiple nodes in the cluster, removing any single point of failure from the storage. As a unified storage platform, Scale’s storage clusters also provide for simultaneous SAN and NAS services from a single pool of storage, providing flexibility for future use and more convenience for IT administrators.

We kept coming back to the fact that Scale’s technology really is revolutionary. The ICOS technology was the difference-maker for us in that we had drive redundancy and node redundancy within our clusters. As we kept comparing and contrasting to controller-based storage; the technology is very different than anything else out there.

In January 2011, Speidel and his team implemented the Scale solution, which included one storage cluster at City Hall and another at the police department. The package included training for staff and the five years of support that Speidel wanted to ensure they could use the storage regardless of what happened to their budget in the upcoming years. The previous MSA systems were repurposed for disk-based backup, improving file recovery time for customers while decreasing time demand on staff. Today the City of Terre Haute has 35 virtualized servers, two VMware clusters of servers and eight VMware nodes between the two sites.

Benefits

Using replication built into the storage cluster, Speidel has been able to beat recovery time and recovery point objectives and develop a plan for maintaining public safety in the event of natural disasters. Scale’s architecture has also provided the city a realistic, affordable storage growth plan for the future that will avoid the same roadblocks Speidel hit with his previous storage. Scale Computing has helped Speidel make the most of his budget. Because the Scale solution consumed a smaller footprint in the data center than their previous storage, Speidel avoided the additional expenses of a third data rack, redundant power and UPS capacity. The cost savings from this project is enabling him to now consider best-of-breed backup solutions to pair with his Scale storage, which will still cost less than other bundled storage and backup solutions.