SQL Server Data Services: What Is It?
My colleague, senior editor, Kathleen Richards was at MIX08 in Las Vegas two weeks ago. Suffice to say she received a brain dump on Silverlight and a number of other big announcements there. When she called from Vegas, my first question was, how's Silverlight? It was quite impressive, she said, adding that there was another big surprise: Microsoft launched at MIX08 SQL Server Data Services, or SSDS.
"Is it a hosted version of SQL Server," I asked her? It's a cloud-based repository service that will offer APIs that enable storage in the cloud, she said (which she also wrote in her subsequent report).
I suspect there are still a lot of developers wondering what SSDS is and what it means to the future of SQL Server and data-driven applications. According to a Microsoft FAQ, it a storage-based service that will enable quick provisioning of on-demand, data-driven and mashup applications and other Web 2.0 apps.
Kate's report from MIX, which is in the current issue of Redmond Developer News, provides a great synopsis of what SSDS is about. Here are some key points from her report:
SQL Server Data Services uses a simple schema based on a model that's defined by three levels: authority (location of the data), containers (unit of partitioning the data) and entities. SQL Server is "the DNA of the service," says Tudor Toma, a group product manager at Microsoft. Currently, it supports SQL Server 2005.
The objective is to provide a low barrier to entry, Toma explains, and then based on developer and partner feedback, Microsoft will add more functionality so that the service is closer to what SQL Server offers today. Toma's group is also considering on-premise service APIs.
Queries are made using a LINQ syntax from C#. "That's how you express the query," says Soumitra Sengupta, a Microsoft program manager, "but you can consume it from PHP, from Ruby on Rails, from whatever environment-it's completely based on REST verbs."
Like Astoria, SQL Server Data Services will support ATOM and the ATOM Publishing Protocol. Microsoft is not going forward with Web3S.
SQL Server Data Services will be a closed beta program until the second half of this year, says Toma. Updates will be released every few months, and the service is expected in the first half of 2009. Interested developers can register to participate in an invitation-only beta on the homepage.
Thanks Kate for that very insightful report. Some other points that bear noting, excerpted from Microsoft's FAQ on SSDS:
Will SSDS support everything in SQL Server?
SSDS will initially only expose a small subset of what is in SQL Server. Microsoft says it plans to extend the capabilities of SSDS over time with such features as binary large object columns, full text search, and richer data types.
How will developers and customers use SSDS?
It is designed to let developers rapidly run on-demand applications, while end users will be able to query and modify any amount data. SSDS will support REST and SOAP interfaces, which are designed to work with any Internet-development toolkit. XML is the primary interface.
What might SSDS be used for?
Data archival, reference data such as catalogs, business functions such as HR services (ie. healthcare records management), and Web facing applications including social networking and picture sharing.
While it remains to be seen whether SSDS will catch on, it is a safe bet that this will not be appealing to those running mission-critical or sensitive data any time soon. Also the proof will be in other details, including proven performance, security, reliability and, of course, price. Since Microsoft is targeting SMBs, the implication is this will be less expensive than deploying in-house, data-driven applications. While the devil will be in the details, SSDS no doubt will be worth watching.
Will this be a flash in the pan, or will SSDS redefine data-driven applications? Please share your thoughts.
Posted by Jeffrey Schwartz on 03/19/2008