It's no secret vendors such as Microsoft and Amazon.com have placed a bet on providing databases in the cloud. Still, a number of challenges remain before enterprises can begin adopting database-as-a-service offerings in a big way-notably concerns over security, vendor pricing models and performance in the case of large application workloads.
Despite these concerns, given the potential for cost savings in the right situation and the reduction of IT administrative burdens, cloud databases should have a place in enterprise DBMS strategies, according to a new report by Forrester Research.
The report, entitled "Database-as-a-Service Explodes on the Scene," advises businesses to take a slow, steady approach to DAAS (database as a service) offerings. Enterprises should start with small, new, non-mission-critical Web applications with structured data that can use the SOAP (Simple Object Access Protocol) and REST protocols. From there, Forrester recommends moving on to mashups that support Web 2.0 applications and mobile applications before ultimately moving on to existing applications that require customization.
"DAAS offers a great opportunity to move some of the non-mission-critical, new Web 2.0 applications, departmental applications and for [small and midsize businesses], focusing on saving money," said Noel Yuhanna, the author of the report, in an interview with eWEEK.
Already, a number of vendors have begun to push databases into the cloud, notably Microsoft with its SQL Server Data Services offering, but also companies such as EnterpriseDB and Amazon.com with its SimpleDB service. Still, there has been a healthy dose of skepticism about whether or not DAAS will make sense for enterprises dealing with large amounts of data and heavy application workloads.
During the next two to three years, Yuhanna predicts there will be improvements around data security, performance and availability that will allow DAAS providers to support more mission-critical and complex applications.
In the report, he recommends that enterprises avoid integration between DAAS and on-premises data for now, as that requires opening special ports, instituting additional security measures and ensuring metadata integration. He also advises businesses to initially store nonprivate data in the DAAS platform and insist on data-at-rest encryption, auditing and secure network connections.
"When database technology first came on the scene some 40 years ago, no one wanted to take their mission-critical applications to such a technology-COBOL, files and paper records were the trusted and reliable approach, but over time that changed," Yuhanna said. "Similarly, DAAS adoption will take time, becoming mainstream in about three to four years when it can support complex and mission-critical applications."