Data Access discussed at Xcalia

This article introduces the EntityBag, as a way to export Entities out of their persistence context. Exactly what JDO and JPA also proposed with attach/detach. This is also not too far from SDO.

Good article seen on the Visual Studio Magazine.

It introduces some new features of the recently released new version (VS 2008 SP1).

Good list of links at the end of the article.

ADO.Net Data Services is going in the REST / Atom direction. Tenants of this approach claim it is simpler than SOAP. When we look at the tools and framework provided by Microsoft we can see that REST is actually hidden by APIs and LINQ. All the REST statements are computed and issued by the framework. There is almost no difference with what an SDO / DAS client programmer will experience.

This other article about the Entity Framework, is also of interest. Related: Microsoft docs about the Entity Framework.

http://www.databasecolumn.com/2008/05/cloud-and-bi.html

The author envisions some impacts of cloud computing on the BI software industry:

• easier evaluation (nothing to download, no machine / space / skills required to install, no configuration, nothing to uninstall)
• more short-term ad-hoc analysis
• more BI projects because funding is easier and less resources required
• extension of data warehouse projects to medium size companies
• analytic SaaS market will grow quickly

Well, maybe it won’t be so easy, it will be interesting to track real projects going in this direction and how they succeed/fail.

Obviously this “BI over the cloud” requires a new kind of database technology (database column is linked  with Vertica):

• Share-nothing architecture, to quickly asborb peaks in on-demand analysis.
• Aggressive data compression to reduce storage costs. ==> Agreed, but what about performance?
• Automatic grid replication and failover ==> The cloud must prove it is at least as robust as an internal installation.

HBase, yet another alternate distributed column-oriented database technology, inspired by Google’s BigTable. It is hosted within the Apache Hadoop project (MapReduce implementation).

InfoQ article about HBase. 

Interesting excerpt about Google’s App Engine:

However, as a lot of people have noted since the announcement of App Engine, there’s a big difference between owning your infrastructure and renting it. It’s probably a very good thing for you when you are small, but as soon as you reach a surprisingly low threshold, you’re better off hosting it yourself.

Also seen on the db4o community, but it could be of interest to everybody: http://developer.db4o.com/blogs/carl/archive/2008/05/02/linq-for-java.aspx

Old post about Java and LINQ.

Carl (db4o’s CTO) would really like to start this initiative, as an open source development, as a JSR and as an Eclipse plugin. A Google group has already been started.

Please note IBM PureQuery (as part of their Data Studio) has some sort of limited support for LINQ. USed to be known as JLINQ.

Quaere, another limited LINQ for Java implementation, hosted on CodeHaus.

Jonathan Bruce’s post about Quaere.

I’ve also seen that one: JoSQL.

Charlie Calvert’s links on LINQ. At Xcalia, we enjoyed working with Charlie’s team in the TAP program in 2007, mostly around LINQ.

Interesting to see that more and more vendors are embracing LINQ.

See these posts from the db4o community:

• http://developer.db4o.com/blogs/kudos/archive/2008/04/30/kudos-linq-for-db4o-is-extremely-easy-to-use-and-very-powerful.aspx
• http://developer.db4o.com/blogs/community/archive/2008/04/30/an-introduction-to-linq-for-db4o-by-edwin-vermeer.aspx

LINQ for db4o.