eek! Update: Try out the SQL Azure Database CTP today!
This one is for the architects! I’ve been catching up on the nosql phenomenon (and think that Windows Azure Storage should do just fine for “nosql” types). But for those of us who cherish relational databases (especially SQL Server) and want it in the cloud at scale, we’ll have SQL Data Services real soon. Oops, I’m so last week. I meant we’ll have SQL Azure Database real soon.
What follows are some links to get you up to speed on SQL Azure (if you don’t have the time to review the videos, hopefully my guide posts will help). Join the mailing list to get notified when they’ll have bits to test.
1. Get to know SQL Azure (nee SQL Data Services nee SQL Server Data Services) on the team blog and at the MSDN Developer center.
2. But does it scale? Read the Exchange Hosted Archive’s story on the team blog.
3. SQL Azure in a nutshell, watch a ~25 minute video to learn:
notion of the Data Hub (@ 2:40), does it work with my SQL Management Studio, SSIS and bcp? (answer @ 5:00), what makes it all work TDS (@ 6:05) , TSQL what works and what doesn’t in “Version 1” (aka V1) (@ 8:00), DBAs Do I have a job? (@ 10:40), Physical vs Logical Administration (@ 12:00), “Taking out the Garbage” (@ 12:45), Frictionless getting started (@ 14:40), Porting existing applications – “Code Near” vs. “Code Far” (@15:30), Suggested Partitioning pattern (@ 19:00), @ 21:00 some closing thoughts including developer portal improvements, potential general availability and more V1 limits such as no link servers, and a “nosql” type question: Azure Storage vs. Relational SQL Azure.
4. If you want more, my recommendation is this well done Channel9 video entitled The Future of SQL Data Services with Nigel Ellis. It is longish, around 56 minutes, but worth the time investment if you plan to architect for this new world. This is an “Inside SQL Azure” conversation that shows how the service is architected and works under the covers.
0:00 - 6:50 Brief History of the product, “CloudDB”, and the former “ACE Model” API. 6:50 – 11:40 The how and whys of data replicas, leader election algorithm, HA out of the box vs a VM model. 11:40 – 15:45 The architecture of the service (load balancer, gateways, data storage), how TDS fits in (and opens the service to the world of TDS Clients – Ruby, Java, .NET, Linux, etc.), statefull vs. stateless gateway tier. 15:45 – 20:25 Challenges of statefull TDS protocol, how databases are provisioned, multi-tenancy, load balancing, resource governance, the SQL Server codebase and its connection to this service. 20:25 – 28:50 Application Patterns and the life of the developer, code near vs. code far, SLAs, timeline and march to general availability, logical vs. physical, policy based management. 28:50 – 37:10 Details of the “Version 1” (aka V1) service including constraints – no cross database queries, use database changes. The methodologies used to come up with the common patterns and compatibility between on-premise SQL Server and the service, the need for an “upgrade advisor”. 37:10 – 43:30 Migration from ACE model – think ADO.NET Data Services and Entity Framework or WCF. 43:30 – 46:00 Guidance for you! Stay away from: cross db operations, switching db context, distributed transactions, Service Broker, SQLCLR (out of scope for V1). 46:00 – 56:00 The Future and Vision, watch this one if this is the only one you watch! – benefits of Azure, ubiquity of cloud, “data hub”.
For a majority of applications, SQL Azure may be a great fit…hey, you know how to connect and program it already (and now with Internet scale).
P.S. If you followed that link on nosql, you’ll see a link to Cliff Moon’s session on Dynomite at Powerset. Powerset?