Comments on: London Stock Exchange Lesson

By: Martin

Martin — Tue, 20 Oct 2009 09:40:48 +0000

My main reason for this post was fact that a years ago I read Microsoft Success Story about LSE – http://www.microsoft.com/uk/getthefacts/lse.mspx and was very surprised. A few years later this system is replaced by cheaper with better performance system built on top of Linux platform (from company that is specialised on this kind of software). That is it. I just do not like strong words (MS success story). It is true, that you can build good or bad software on whatever platform for less or more money.

By: Ron

Ron — Sun, 18 Oct 2009 21:55:20 +0000

Ok, honestly, how many database systems with 900+ trx/sec have you seen recently? For example in payment card industry (like e.g. Postilion)? What I think is that you’ve read one blog about .net&mssql failure in LSE and write your own “comment” without any knowledge how it works, what is the true cause of the failure and without any added value. It is not about what’s happened and why mssql server is worse than its better competitors but about your approach. You don’t know how the LSE failure was divided between database and .NET application layer but you “certainly” know that you can’t achieve “99,8% on top MS SQL 2000″ and “Windows platform is worse than Linux”. I am sorry but this is crap.

By: Martin

Martin — Sun, 18 Oct 2009 09:59:16 +0000

Well maybe you are right. I was working as a software engineer for enterprise systems (telco and mainly financial systems) only for last let say six years. I am using MS technologies from MS SQL 7. I think that it is honest to say that MS SQL 2005 is good server, but it was not so good with previous versions. I believe it is possible to build such system using MS technologies, but I am just wondering why LSE moved from so good solution to other solution that is cheaper and has better performance.

By: Ron

Ron — Sat, 17 Oct 2009 19:35:26 +0000

It seems that you know completely nothing about enterprise systems availability. 99.8% availability can be definitely reached by any MS SQL platform it was ever released. This availability is about 17 hours downtime per year which is very well possible. There are proven deployments where there is synchronization on disk array level which allows active-active scenario. This kind of scenario allows under certain circumstances (redundat other hw) availability near to five nines (~5 minutes/year).