New-age Transactional Systems - Not Your Grandpa's OLTP
John Hugg discusses high volume transaction processing applications with high and low frequency profiles, and how VoltDB can be used for that purpose.
The content has been bookmarked!
There was an error bookmarking this content! Please retry.
Posted by Al Tenhundfeld on Jan 17, 2009
Versioning database schema along with your .NET code is essential for managing volatile codebases especially when employing continuous integration. Many teams in the .NET space use handwritten scripts or schema comparison tools. Ruby on Rails accomplishes this with a popular solution of abstracting DDL SQL into Ruby commands called migrations.
The following Rails migration, written in Ruby, defines the actions for creating and dropping a Users table in a database:

Using the RikMigrations library, similar code can be written in C#:

The important concept to understand is that all of the data definition language defining the database schema has been abstracted and moved into the application code. This has several advantages:
Migrations are still not widely used within the .NET community. Unfamiliarity with the approach accounts for much of that, but there are some valid arguments against migrations. Many .NET teams make wide use of database stored procedures. For systems using stored procedures, a versioned script approach might work better, and it is likely platform neutrality is not a concern. Also, for large applications that have databases managed by DBA's, moving DDL into .NET code may not be an option.
There are two .NET migration libraries that have growing communities: RikMigrations (code) and Migrator.NET .
RikMigrations has been the more popular library, supporting a more fluent interface and a command line interface. However, the main developer stopped contributing to it in the middle of last year. Migrator.NET is growing in popularity and maturing quickly with a new fluent interface and automation integration. Both are small open source projects that could use more support from the developer community.
Justin Etheredge, C# MVP, has a written a useful tutorial on getting started with RikMigrations, including useful pointers on configuration.
Why NoSQL? A primer on Managing the Transition from RDBMS to NoSQL
A Guide to Branching and Merging Patterns
SCM best practices for multiple processes, releases & distributed teams
DBDeploy is a similar tool that enables this with pure SQL scripts. I originally like the code wrapping and the platform independence created by a framework like the ones you mention. But most of them falls short when it comes to doing migrations that need to change existing data in the database. I haven't really looked into the frameworks you mention, but we found that some of the stuff we needed to do with out data when refactoring was quite heavy, even when it comes to SQL.
See blog.f12.no/wp/2009/01/03/migrations-for-java/ for a similar article about DBDeploy and focused on Java. I think there is a .NET version of DBDeploy too.
Also worth looking at is the tarantino project - Database change management; nant task to handle database updates, it does not abstract away the sql but I have found it to be a very good tool for managing database changes
code.google.com/p/tarantino/
Since its C# code, there's nothing stopping you from using a helper function to abstract a connection to the provider, and executing SQL.
When Migration was gaining popularity and there was no mature versions of it for .NET, I took it upon myself of writing one. It took about 2 days, but I took a shortcut: I used SMO, which tied it to SQL Server, with some helper methods (it predated .NET 3.5, today using extension methods would make it vastly cleaner) to duplicate most of the functionality (improved on it in a few cases too, hehe).
So for some stuff that was easier done in SQL, we just executed scripts, or inline SQL if it was minimal (though using scripts can let you have platform independance: if you have different folders for different databases...sure you need to duplicate the work, but it shouldn't happen too often).
So point is, in the end, its still C# code, and you can do everything, you're not tied to using the migration framework alone... That is actually why we went the C# route in the first place: We could easily execute SQL from C#. Executing C# from SQL is trickier unless you're already investing in CLR within your database server.
I haven't heard of DBDeploy. Thanks for bringing it to my attention.
RikMigration does support changing existing data. In fact, it does so by using anonymous types, which is a pretty cool idea in my opinion. In the comments of Justin's post, one of the RikMigrations developers gives an example.
Building on my sample from above, you could issue these commands:
usersTable.Insert(new{ ID = 1, first_name = "Al"});
usersTable.Insert(new{ ID = 2, first_name = "Anders"});
Cool. I think for a lot of shops that already have a collection of scripts, moving to a migration approach may be more than they want to tackle, but organizing their scripts into a framework like tarantino or DBDeploy may be very helpful.
Thanks for bringing this to my attention.
octalforty Wizardby is kinda similar, but it uses a special language (call it a DSL if you like), which is much expressive than C# is for this particular purpose. It can also generate "downgrade" migrations automatically and has a clever compiler which allows for some type inference and intelligent naming of FK references/indexes.
See code.google.com/p/octalforty-wizardby/ for more on that.
John Hugg discusses high volume transaction processing applications with high and low frequency profiles, and how VoltDB can be used for that purpose.
Kevlin Henney examines code samples to see what can be learned from them starting from the premise that one won’t write great code unless he knows how to read it.
Jason Ayers share the observations he made watching a team of developers collaborating in real time on the same code base, pushing XP, pair programming and continuous integration to their extremes.
Michael Snoyman presents Yesod, a web framework written in Haskell and containing a web server, templating, ORM, libraries (templating, gravatar, etc.).
Richard Kreuter and Kyle Banker on how to avoid classical RDBMS transactional systems by using compensation mechanisms, transactional messaging or transactional procedures.
Attila Szegedi talks about performance tuning Java and Scala programs at Twitter: how to approach GC problems, the importance of asynchronous I/O, when to use MySQL/Cassandra/Redis, and much more.
One category of risk that project teams need to ensure they address is business value failure – delivering a product that fails to provide value for the business investor.
InfoQ spoke to the authors of Software Systems Architecture on a couple of new topics, the System Context viewpoint and Agile, which have been added to the second edition.
6 comments
Watch Thread Reply