Comments on: NoSQL and the End of Object Relational Models http://willcodeforcoffee.com/2010/03/02/nosql-and-the-end-of-object-relational-models/ The personal programming blog of Eric Hoff. Tue, 02 Mar 2010 21:41:58 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: eric http://willcodeforcoffee.com/2010/03/02/nosql-and-the-end-of-object-relational-models/comment-page-1/#comment-481 eric Tue, 02 Mar 2010 21:41:58 +0000 http://willcodeforcoffee.com/?p=121#comment-481 Thanks for the comment Bart! I suppose one of the points that I should have touched was the fact most developers and architects are experienced and comfortable with the the Relational Model and they are <strong>not</strong> comfortable with a NoSQL model. It did not surprise me at all that Microsoft's customers wanted an RDB because Relational is proven. Azure isn't (only because it is still so new), so it might be risky. Learning a new technology is fun for one developer, hard for a team - and even harder when migrating an existing project. Having to switch only one technology at a time reduces some risk. It will take baby steps to move into the Cloud just like it will take baby steps to move into a NoSQL domain. I still think the two will be linked. Personally, I think that Azure is one of the best options available because of the choices. I also think that offering SQL Azure was a good decision. At the moment, it is hard to say what will affect a customer's or architect's about using Azure's services and which services they use, but as comfort with Cloud services increases I think we might see more Architects switching to Azure Storage. Thanks for the comment Bart!
I suppose one of the points that I should have touched was the fact most developers and architects are experienced and comfortable with the the Relational Model and they are not comfortable with a NoSQL model. It did not surprise me at all that Microsoft’s customers wanted an RDB because Relational is proven. Azure isn’t (only because it is still so new), so it might be risky. Learning a new technology is fun for one developer, hard for a team – and even harder when migrating an existing project. Having to switch only one technology at a time reduces some risk.
It will take baby steps to move into the Cloud just like it will take baby steps to move into a NoSQL domain. I still think the two will be linked.
Personally, I think that Azure is one of the best options available because of the choices. I also think that offering SQL Azure was a good decision. At the moment, it is hard to say what will affect a customer’s or architect’s about using Azure’s services and which services they use, but as comfort with Cloud services increases I think we might see more Architects switching to Azure Storage.

]]> By: Bart Robertson http://willcodeforcoffee.com/2010/03/02/nosql-and-the-end-of-object-relational-models/comment-page-1/#comment-480 Bart Robertson Tue, 02 Mar 2010 20:44:32 +0000 http://willcodeforcoffee.com/?p=121#comment-480 Disclaimer: The opinions are my own and not necessarily those of my employer, Microsoft. Having spent the last year plus working heavily with Windows Azure and SQL Azure, there are definitely some benefits to Azure storage, particularly in the prototype / proof-of-concept phase. Azure Storage also scales well, so if you’re building a “mega-service”, Azure Storage is a great choice. As an architect, you should definitely evaluate and understand the NoSQL options and their trade-offs vs. the relational DB options. The SQL Azure team made a very difficult decision last spring to go away from the NoSQL approach and, instead, provide a relational database in the cloud. The feedback from our customers was loud and clear – we want our relational database. While it’s easy to get excited about the promise of NoSQL, RDBs have had decades to mature, and you very quickly start to miss features that you take for granted with a modern RDB. As SQL Azure matures, the features and capabilities will continue to increase, as will the value of an RDB in the cloud. The great news about the Windows Azure platform is that it provides choices that allow you to make the best decisions given the requirements, so it’s possible to leverage the best of on-premise, the cloud, NoSQL and a full-featured relational database like SQL Server. Disclaimer: The opinions are my own and not necessarily those of my employer, Microsoft.

Having spent the last year plus working heavily with Windows Azure and SQL Azure, there are definitely some benefits to Azure storage, particularly in the prototype / proof-of-concept phase. Azure Storage also scales well, so if you’re building a “mega-service”, Azure Storage is a great choice. As an architect, you should definitely evaluate and understand the NoSQL options and their trade-offs vs. the relational DB options.

The SQL Azure team made a very difficult decision last spring to go away from the NoSQL approach and, instead, provide a relational database in the cloud. The feedback from our customers was loud and clear – we want our relational database. While it’s easy to get excited about the promise of NoSQL, RDBs have had decades to mature, and you very quickly start to miss features that you take for granted with a modern RDB. As SQL Azure matures, the features and capabilities will continue to increase, as will the value of an RDB in the cloud. The great news about the Windows Azure platform is that it provides choices that allow you to make the best decisions given the requirements, so it’s possible to leverage the best of on-premise, the cloud, NoSQL and a full-featured relational database like SQL Server.

]]>