The Metropolitan Sewer District of Greater Cincinnati, Ohio, (MSD-GC) has relegated EMC Corp. to secondary storage and promoted Hitachi Data Systems Inc. (HDS) to primary storage using some nifty third-party storage replication software from Topio Inc. The changes allowed the department to shave 18% off its storage hardware spending, slash its administration costs and perform migration jobs with next to no downtime, according to district...
MSD-GC supports 600 users at seven locations and has been on a disaster recovery kick. The department bought an EMC Clariion CX600 three years ago and then spent a year evaluating a second purchase for disaster recovery. MSD-GC looked at EMC's MirrorView storage replication product with a mind to buying a second Clariion and replicating to this device.
"With EMC we had to have the exact same hardware at the recovery site, it had to be a homogenous storage area network (SAN), which was very expensive," noted Don Sander, IT systems analyst at MSD-GC. And the department hadn't exactly been bowled over with the management of the Clariion.
Configuring the Clariion CX600 is like trying to understand "Chinese arithmetic," according to Marty Hubbard, IT systems analyst at MSD-GC. "Working out what goes to what is impossible; the naming conventions don't make sense." He added that in 2003, when the company initially purchased its Clariion, EMC and its services partner Unisys took over the implementation. "We wanted to have more input into the naming conventions when setting everything up, but that was impossible," Hubbard claimed.
In late 2004, the company changed directions and bought an HDS 8595 array and heterogeneous replication software from Topio, an HDS partner, for its secondary storage site. The HDS SAN is housed in another building on the 72-acre site in Hamilton County. The initial plan was to replicate the EMC Clariion SAN, holding about 8.6 terabytes of data, to the HDS SAN strictly for disaster recovery purposes.
Then MSD-GC realized that once the data was replicated, using Topio enabled it to switch the sites around to make HDS its primary storage. The HDS box performed faster, was easier to configure and came in at 18% cheaper than another EMC Clariion, according to Sander; he wasn't able to disclose performance or pricing details. The ability to support heterogeneous storage allowed MSD-GC to deploy HDS at its secondary site, lowering its hardware costs. But the reduction in management time was the real kicker.
"With Topio, you just set up, let it get synchronized and then forget about it," Sander said. Topio's Data Protection Suite (TDPS) takes the brunt of the processing off the primary storage by performing the heavy lifting on a dedicated server at the remote site. TDPS requires an agent on the host at the primary site that intercepts writes, time-stamps them and sends them over the wire to the recovery server where they are reassembled in order. Moreover, TDPS only copies changed bits, reducing the time to perform the replication and the amount of storage needed at the recovery site.
MSD-GC tells of a time when it had to migrate 700 GB of data to a new server, which took down an entire department for a day. "Fifteen-to-20 people were doing nothing," according to Sander. "We had people yelling at us." He said if they'd had Topio at that point, nobody would have blinked an eye. "We could have done it in a second."
Topio TDPS 3.0
This week Topio has launched TDPS 3.0 adding a one-to-many replication feature. This capability allows source data to be simultaneously replicated to multiple destinations in an organization, providing multisite business continuity, local and remote data protection, or continuous data sharing for applications, such as development and test or analytics.
Additional features in version 3.0 include management control options that allow centralized operational control of distributed replication services. With TDPS 3.0, authorized administrators can remotely perform tasks that in conventional products require an on-site operator. New quality of service (QoS) features enable prioritization of service levels and assure key applications will receive appropriate resources first. Topio has also beefed up its management interface with historical reporting, providing users increased perspective for trend analysis, resource and service level optimization, trouble shooting, and root-cause analysis.
MSD-GC would like to see Topio add a universal restore capability to its software, meaning that users could buy different servers for the recovery site than what's deployed at the primary site. At the moment, they have to be the same. "We like to be able to use old servers in disaster recovery and then bring in new ones at our primary location," Sander said. "It would be nice if Topio had that instead of us having to buy a separate product." This is only an issue on the operating system side, as Topio supports a mixture of any storage hardware.
It's worth noting that since MSD-GC's implementation, EMC claims to have simplified the management of the Clariion and is enabling users to install and configure the product themselves. The company also acquired long-distance, heterogeneous replication startup Kashya Inc. recently, adding a well used piece of technology to its replication arsenal.
Other vendors with products that perform storage replication like Topio's include: XOsoft Inc. (owned by CA Inc.), Constant Data Inc., FalconStor Software Inc., Kashya Inc. (owned by EMC), NSI Software Inc. and Sanrad Inc.