Virtualized Storage

By the power of disk-based data protection and storage virtualization technology, renowned energy service provider Champion Technologies eliminated the backup window for its SAP business system, saving a million dollars in IT overhead and achieving global business continuity -- even during a massive hurricane. At the helm of this project was Champion VP/CIO Roland Etcheverry, who describes his experiences.

When one of the largest hurricanes in U.S. history struck Texas in 2008, it wreaked havoc and caused billions of dollars worth of damage to many businesses. Because of our backup and disaster recovery (DR) infrastructure, we were thankfully spared from this devastation, secure that our data was fully protected and easily recoverable.

Headquartered in Houston, Texas, Champion Technologies is one of the fastest growing specialty chemical companies. We have a hosted collocation facility in Scottsdale, Ariz., and branch offices and manufacturing sites spread globally across the U.S., Latin America, Europe, Asia Pacific and Russia.

Like many businesses, we rely on SAP ERP and Microsoft Exchange systems for 24x7 operations and messaging services. Protecting the data on these systems is vital to our success and survival. Several days of downtime could seriously affect our manufacturing and shipping processes. As the CIO of Champion, it is my responsibility to maintain ongoing operations and data availability, regardless of the circumstances.

Making the Switch to Virtualized Storage

Our IT team originally used third-party tape-based backup and DR services for data protection. This was reliable, but problematic in that it forced critical systems to be down for extended periods of time. A typical backup would take over 14 hours, and recovery would take more than two weeks, during which time our business critical systems would need to be offline. As the company and its data grew, this methodology became increasingly unacceptable.

In an effort to reduce backup and recovery times, we set out to build a streamlined, high-performance, in-house DR infrastructure. Implementing a cost-effective SAN was critical, because it would eliminate the need for each server to have its own direct attached storage. After careful research and testing, we purchased the FalconStor Network Storage Server (NSS) solution to provision SAN storage to over 225 servers at our headquarters. Its storage virtualization and provisioning technology offers a multitude of capabilities, including snapshots, data replication, synchronous mirroring, and active/active failover. In addition, snapshot images are provided to support our SAP applications, Fibre Channel (FC) connectivity support for our Dot Hill storage arrays, iSCSI support, and high performance for writes and rewrites of the SAP/ERP system.

We installed two appliances at the data center in a high availability (HA) active/active pair for redundancy. Another appliance was installed at the DR site, acting a replication target and a gateway to our Dot Hill FC storage. Application-aware snapshot agents ensure transactional consistency of data, which enables us to recover SAP and Microsoft Exchange data without having to perform bare metal restore. We take scheduled hourly snapshots, which are replicated over IP in a WAN-optimized fashion, minimizing bandwidth requirements and, in turn, minimizing our backup window. Delta snapshot images of data offer point-in-time, ready-to-use copies in case of data loss or system failure.

Worth a Million Bucks... and Then Some

If a failure or disaster occurs, we can simply roll back data to a previously good state and resume operations. Applications can be restored in as little as two hours, rather than two-weeks. With this new system, we no longer have to question whether the data is protected, or whether we can recover data quickly enough to meet internal and external SLAs.

Our new DR system was put to the test, and its true strength revealed, in a real-life failover scenario last year. When meteorologists warned about Hurricane Ike hitting the Houston area, we knew that it could potentially shut down our company's headquarters for several days or weeks. In anticipation of this storm, we executed a failover of the business systems to the DR site. To our delight and relief, the entire failover and failback process only took approximately four hours.

When the headquarters building became available five days after the storm passed, the business systems were failed back to the Houston data center. Throughout the outage, users in global offices continued to work without a hitch. From a business system perspective, Hurricane Ike was a non-event.

Some say that you cannot put a price on the value of data availability, business continuity, and peace-of-mind. This is only partially true. To date, we have saved a million dollars in hardware purchases and storage management overhead. We anticipate additional savings over time as we move towards a completely tapeless protection model, and as we avert disasters in the future.

Interested in information related to this topic? Subscribe to our Information Technology eNewsletter.