Interesting article, you've obviously put some thought into how to architect a robust logging environment. I wish someone had done the same in my environment, instead hundreds of programs use centralized logging to single table. Now, I'm all for centralized and standardized logging however, if this one table on this one database on this one clustered database server isn't available--everything stops. So, logging must be asynch and not cause the application to fail. I think your solution accomplishes this to some extent by using Service Broker.

One suggestion, you might want to take a look at a more featured solution that also uses Service Broker and the open source Log4Net framework is described in this Dr Dobb's article:

Enterprise Application Logging Using the SQL Server 2005 Service Broker