eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Mitchell International Inc. needed to find a way to archive its data—while keeping it easily accessible on demand—to avoid a storage wreck.
Mitchell, based in San Diego, helps facilitate the claims process for insurance companies and auto repair shops by automating the procurement of services such as collision repair, medical claims and glass replacement. With explosive database growth, though, Mitchell wanted to have more flexibility in its archival process as well as to control storage costs.
Last summer, the company decided to deploy Princeton Softech Inc.s Archive for Server software (also known as an Active Archive solution). Active Archive is an ILM (information lifecycle management) solution that has given Mitchell more flexibility in the backup process.
“We serve insurance carriers, and they are getting more demanding regarding how they want data pulled for them, and this solution enabled us to more actively search archives,” said Dave Dullaghan, director of database architecture at Mitchell. “Having this type of solution really enables us to be more proactive with our offerings.”
During eWEEK Labs exclusive on-site visit at Mitchell headquarters, we were impressed with the companys emphasis on data protection and with its optimization of storage resources. By deploying Princeton Softechs Active Archive software, Mitchell has been able to develop an ILM strategy that not only improves application performance and data availability but also significantly reduces time to market for revenue-generating services.
As mushrooming database growth degrades database and application performance while increasing storage and operational costs, enterprises like Mitchell are looking for the best way to manage and store data through the information life cycle—or risk facing a loss of productivity and revenue opportunities.
Mitchell isnt the only company with growing storage concerns. Last year, Meta Group Inc. predicted that compounded annual growth for databases will exceed 125 percent this year. The research company also projects that database archiving will become a best practice in enterprises by 2007.
Founded in 1946 as a creator of collision-estimate books, Mitchell has revolutionized the automobile insurance claims process with its e-business solutions. The companys Internet-based product offerings have enabled insurers to adopt online processing and management of collision repair claims and include capabilities such as online digital imaging and online estimating for body shops.
Each month, Mitchell facilitates several million electronic transactions for its more than 16,000 business partners.
The number of transactions Mitchell processes is one reason for the companys rapidly growing storage needs. As the data accumulates, the relationships in the data grow increasingly complex, particularly as the company offers more feature functionality to its clients. As a result, Dullaghan wanted a way to have more control over predicting the optimization of his databases.
In addition, while client information may be rarely accessed, files associated with an insurance claim must be kept available and maintained for a year and must be accessible after that for an insurance claim to be closed. Mitchell was already performing backups of databases, but Dullaghan wanted to supplement those backups while providing flexibility in searching archives.
“It became much more of a hassle to be restoring a single estimate for one customer,” Dullaghan said. “We needed to find something that would allow us to get into the indexed archive file quickly and pull that archive back in an efficient manner.”
Next page: Evaluating the options.
Page Two
Last summer, Dullaghan began looking at ILM solutions from Princeton Softech, EMC Corp.s Legato Software division and Veritas Software Corp. Dullaghan also considered building the archival solution in-house.
In the end, Dullaghan turned to Princeton Softechs Active Archive and began implementation in January. (He declined to provide cost estimates.)
“We didnt feel like the other products came close to our problem domain,” Dullaghan said. “Considering the complexity of the [data] relationships we have, we wanted the ability to link multiple databases together and really be able to actively manage the number of records in our database.”
Using Active Archive, Dullaghan can offload older database data from Mitchells EMC Symmetrix DMX Series system, which is used for primary storage. The system archives data onto an EMC Centera NAS (network-attached storage) device to maintain its referential integrity. The data is kept in a searchable Princeton archive for 12 months and then moved to tape backup.
Dullaghan archives the data and then goes through a deletion phase after confirming the data has been archived. If a restoration is necessary to fulfill a claims request, the restored file is moved into temporary space rather than into a production database.
Mitchell has deployed Active Archive across multiple databases for data warehouses and for large online data stores. Dullaghan estimates that it typically requires about three weeks to go live, including developing the scripts, testing and documenting them, and then deploying them.
Had Mitchell developed in-house the ability to search through multiple archives, Dullaghan estimated it would have taken at least a year.
By employing Active Archive, Dullaghan has freed his developers to concentrate on product functions instead of core capabilities, he said. Requests from insurance carriers that once required more than a month for searches of multiple years now take only a few days, he said.
“Because we can archive multiple databases, we have an advantage in that weve got a logical data set thats consistent and coherent,” Dullaghan said. “What future-proofs [Mitchell] is that Active Archive stores the data definition language with the archive, too; if you make changes to the database in the future, the product will recognize you made a change and will allow us to remap that data.”
In the future, Dullaghan said he wants to automate the archival process after he finds a procedure for comparing the restored data with what is in production to ensure that no changes have been made. In addition, he is looking to link Active Archive with his Legato Software network.
“One of the reasons we do the tape thing is for off-site storage,” Dullaghan said. “Mitchell has a secondary restore facility, which almost gives us quadruple redundancy. Its like having a second insurance policy—when the data is this important, you have to be careful.”
Senior Writer Anne Chen can be reached at anne_chen@ziffdavis.com.
Be sure to add our eWEEK.com developer and Web services news feed to your RSS newsreader or My Yahoo page