Data Access blog : Misc

Quick update re: Data Access blog.

alexbarn — Wed, 12 Jul 2006 00:57:00 GMT

We're putting this blog (Data Access blog) into suspended animation. That doesn't mean we will stop blogging about ADO.NET and data access stuff, or that we'll take down published posts or comments here - it's just means we'll carry on blogging about ADO.NET (current and future releases including info on ADO.NET Entity Framework, EDM, LINQ to Entities, etc) at our shiny new ADO.NET blog and blog about other data access technology related posts at our other team blogs (see below).

If you are subscribed to this blog's feed (there are quite a few of you!), please update you reader / aggregator so it points to the new ADO.NET blog's RSS feed or Atom feed. We can't automagically redirect you, so sorry for the hassle here.

Comments on this blogs.msdn/com/dataaccess blog will be closed, but we have re-published this blog's June and July posts over to the new blog, so you can comment there if you like.

In case you are wondering, here's run down of the various blogs run by the Data Programmability teams:

The Data Programmability team blog (more about data access in general, rather than about a specific technology)
SQL Protocols team blog (discussions related to Microsoft's SQL Server Protocols - Netlibs, TDS and (new for SQL 2005) SOAP. Topics include connections and SQL connectivity)
XML team blog (covers XQuery, XPath, SQLXML, XML Editor, XSLT, XSD, XLinq (LINQ to XML) and more XML goodness)
ADO.NET team blog (ADO.NET and System.Data namepace, posts on current and future releases including info on ADO.NET Entity Framework, EDM, LINQ to Entities, etc)

Also, check out the MSDN Data Access and MSDN XML Developer Centers.

Hope to see you there!

Developers! - Express contest - win $10K

alexbarn — Tue, 14 Mar 2006 21:45:00 GMT

A new worldwide contest for developers has been launched by the folks at Visual Studio Express and SQL Server Express.

Here's how it goes: you think of something cool using Visual Studio Express and SQL Server Express. It could be desktop app, a web app, a mobile app, whatever. Then you submit your cool idea here.

Once the finalists are picked, then you build. The top prize is $10,000 (USD) in cash, another winner will win $1,000 in cash, and all 12 finalists who complete their project will receive $250 in Amazon gift certificates.

There are three judges. Microsoft's Robert Scoble, MAKE's Philip Torrone and a 'Mystery Judge'.

Here's the contest site....good luck!

An Introduction to Data Structures

alexbarn — Mon, 13 Mar 2006 19:53:00 GMT

In November 2003 MSDN Online published a popular six-part series on using data structures in the .NET Framework. In 2005, the first part was updated to take advantage of the new data structures and features available with the .NET Framework version 2.0, and C# 2.0:

"We'll examine both built-in data structures present in the .NET Framework, as well as essential data structures we'll build ourselves. This first part focuses on an introduction to data structures, defining what data structures are, how the efficiency of data structures are analyzed, and why this analysis is important. In this article, we'll also examine two of the most commonly used data structures present in the .NET Framework: the Array and List."

(This article assumes the reader is familiar with C#).

You can read the updated article here.

Conversation with scientist, engineer and database legend Jim Gray

alexbarn — Thu, 09 Mar 2006 22:40:00 GMT

Channel 9 has posted great video interview with Jim Gray, a researcher and manager of Microsoft Research's eScience Group.

"This episode features Jim Gray. He is a "Technical Fellow" in the Scaleable Servers Research Group (Sky Server, Terra Server) and manager of Microsoft's Bay Area Research Center (BARC). Jim has been called a "giant" in the fields of database and transaction processing computer systems. In 1998, Jim was awarded the ACM’s prestigious A.M. Turing Award.

Look out for his answer to the question: "What is your favorite data structure?".

Via Euan Garden.

10 Mistakes Developers Make With Databases

alexbarn — Tue, 07 Mar 2006 02:01:00 GMT

A new article published at Developer.com is worth checking out: Ten of the Biggest Mistakes Developers Make With Databases.

SSL in SQL Server 2005 [Il-Sung Lee]

DataWorksBlog — Sat, 06 Aug 2005 03:48:00 GMT

I often receive questions inquiring about the (channel) encryption capabilities in SQL Server 2005. Like SQL Server 2000, 2005 can use SSL (Secure Socket Layer) to secure transmissions over the wire independent of the network protocol used. However, unlike its predecessor, SQL Server 2005 will always make encryption available, even if the administrator hasn’t provisioned a SSL certificate on the server machine. For this reason, SQL Server 2005 can and will ensure that the login packet is encrypted even if encryption hasn't been explicitly turned on. More precisely, unless either the client or the server requests encryption, the channel will not be encrypted beyond the login packet.

How? When initializing SSL support at startup time, the server will use the following order to load a certificate:
1. Use the certificate that the administrator has specified in the SQL Server Configuration Manager (right-click on "Protocols for <instance>" under "SQL Server 2005 Network Configuration" and the select the "Certificate" tab)
2. If no certificate has been specified, the server will search the machine and user’s certificate store for an appropriate certificate. The following KB article, a explains what the server considers a valid certificate, http://support.microsoft.com/default.aspx?scid=kb;en-us;318605. (Note that the article was written for SQL Server 2000 but the criteria for a valid server certificate are still employed by SQL Server 2005.)
3. If nothing appropriate can be found, the server will generate a self-signed certificate.

Any other implications? The auto-generation of a certificate is a new feature to SQL Server 2005 and is the reason why encryption will be available even though a SSL certificate has not been provisioned on the server machine. This certificate can be used for full channel encryption but with one caveat. If the client is the one requesting encryption, then it will attempt to perform server validation on the certificate to verify the identity of the server machine. This is a prudent security practice but is impossible to do with a self-signed certificate since it hasn't been signed by a trusted root authority. To overcome this problem, the client may specify the "Trust Server Certificate" flag to override the server validation. Alternatively, you can turn encryption on at the server and leave the client encryption flag off to enable channel encryption.

Conclusion: Whether or not you need channel encryption depends entirely upon your system security requirements and performance tolerance. But one of the nice features of SQL Server 2005 is that it will always be available even if you haven’t installed an SSL certificate on the machine (although I still strongly recommend using a certificate signed by a trusted authority whenever possible).

Il-Sung Lee
Program Manager -- SQL Server Protocols

Disclaimer: This posting is provided "AS IS" with no warranties, and confers no rights

Dataworks team at TechEd 2005

DataWorksBlog — Sat, 04 Jun 2005 02:36:00 GMT

How many of you are coming to TechEd 2005 in Orlando? A show of hands, please. Excellent!

The DataWorks team will be at TechEd, delivering sessions, chatting in the cabana and in the booth, and attending various Birds of a Feather sessions.

We'll be involved in the following sessions:
1. Monday @ 5:00 – DAT 330: SQL Server 2005: End-to-End Part 1 (David Sceppa, Euan Gardner and Roger Wolter)
2. Tuesday @ 5:00 – DAT 320: Integrated Innovation: Using ADO.NET 2.0 with SQL Server 2005 (Pablo Castro and David Sceppa)
3. Tuesday @ 3:45 – DAT 421: Client and Middle Tier Caching with SQL Server 2005 (Pablo Castro and Steve Lasker)
4. Thursday @ 8:30 – DAT 420: Roadmap for Native Data Access (Chris Lee)

There will be follow-up discussions in the cabana lounge after each session. The DataWorks team is also delivering presentations in the cabana lounge:
1. Wednesday @ 10:15 – DAT C07: Q&A ADO.NET 2.0 (Pablo Castro and David Sceppa)
2. Wednesday @ 5:30 – DAT C09: ADO.NET 2.0 and the Provider Factory Model (David Sceppa)
3. Thursday @ 5:00 – DAT C13: ADO.NET 2.0 DataSet Enhancements (David Sceppa)

Please stop by and say Hi to us at the booth or the cabana. We look forward to seeing you there!

David Sceppa
Program Manager – ADO.NET Team
Microsoft Corporation

API Changes between CTP [Pablo Castro]

DataWorksBlog — Tue, 05 Apr 2005 17:32:00 GMT

We’ve been seeing several comments on email and blogs lately about changes in the APIs between CTP (Community Tech Preview) releases. I understand the frustration that these changes could cause, so I wanted to jump in and comment on the topic.

As part of being increasingly more open in our development process, product groups at Microsoft such as Visual Studio and SQL Server are using this CTP-centric model as an extension to the traditional Beta programs. Personally I love the CTP model because it allows us to get bits out much more often for our developer community to take a look, start experimenting with the new bits and provide feedback on things they like (or not like :).

Now, this is great but it comes with a cost: releasing CTPs more often means that more of our internal development process is more exposed. During development of large products such as SQL Server or the .NET Framework we experiment, try things out and do some early, tentative implementations. Most of these “experiments” end up being features and we clean them up, add tons of new tests for them, etc. There usually are a few of these experimental features that we decide not to ship, either because we think it won’t be ready for a particular release, because we get strong feedback from our customers saying that the feature is not what was expected, or because the result is not what we expected.

If we decide to remove a feature that we introduced in-cycle, we mark these APIs as [obsolete] for a while, and then we remove them. This gives folks that have applications that were built with these experimental features a chance to see the obsolete warnings and fix their applications before the APIs are removed. At this point is where you see something that you like and just when you got exited about it you realize it’s marked [obsolete] :(

Note that this flexibility during development is key to us. We need to be able to add and remove features so we can see if they’ll live up to the our and your expectations.

So basically the trade-off is: get only a couple of relatively clean Beta version during the development cycle and see a few changes, or get several CTPs during the development cycle (including some betas) and see more of these intermediate features coming in and going away.

Pablo Castro
Program Manager - ADO.NET Team

Disclaimer: This posting is provided "AS IS" with no warranties, and confers