Lextrico

Registrations for the course is now open.  

The topics covered this time around are ‘Helping the Entity Framework play its <ROLE>’, looking at ways of using behavioural roles defined by interfaces in your system to be explicit and achieve high performance results, finer grained separation of concerns and greater configurability between your Domain Model and infrastructure. The second topic will be a practical look at getting started with NServiceBus, including demonstrations taking in messaging patterns such as Publish / Subscribe and full duplex.

While this is something that depending on the complexity and scalability needs of your site one may consider over kill, it is something that takes minimal effort in design to create a system that is innately more scalable. What happens when the data you are querying starts to push out is 5-10Mb of Data to a CSV file?

Option 1

So rather than generate the list of Foo into a CSV at each request why not pre generate the list at pre configured intervals depending on the volatility of the data and save the resultant CSV to a san that can be served up via IIS. In this scenario there is no need for the ASP.NET worker process to get involved and the IIS process can do what it was meant to do serve content.

Advantages

• No need to hit the DB to get the data
• You don’t have to involve the asp.net worker process and let IIS do what its designed to do

Disadvantages

• Your data may be stale up to a preconfigured amount of time (but since its going to a report of sorts it is more than likely stale)

Option 2

So your business requirement absolutely dictates that it is required to be generated and sent as the user requests it. There is no choice, you have but to send it through the response stream. BUT there is another way. Say when this request for the data to be exported VIA an AJAX call to your server the response from the server is a URL to where the file will be output. Now the server it self can then send a message to the export to csv service and request the file be output to the san. The client can then poll the server waiting for a 200 response code, when it does get one it can then redirect the browser to the given URL.

Advantages

• Gives you more or less up to the minute accurate data
• Reduces the amount of work the asp.net worker process has to perform

Disadvantages

• Introduces polling into the design of the client

When designing systems that do not take Scalability into account from the ground up, you generally are missing something. Considering Scalability does not mean that you choose either of these two options all the time. At times it might make sense to think about using the approach suggested in Scott’s article, the key being to always think about know your data and make informed decisions.

My reasons for attending the course again were simple

1. Gauge my progression
2. Hear other peoples thoughts and opinions
3. Spend more time at the feet of the master
4. And of course Learn some more

The beauty of this course is that despite the complexity of the architecture and the concepts the delivery makes the comprehension a breeze.  Paul Batum and Rhys Campbell who attended the course have written up their impressions, as has Simon and we all agree that it is worth its weight in gold. So if you have thought about ever doing this course and can make it to one do so, but if your not able to go to one of upcoming courses I would strongly suggest getting in touch with Udi and working other possible solutions with him. Which is why I guess Simon is owed big thanks for setting up this opportunity.

As for NServiceBus and its comparisons to Nuclear Armament, I can only say that once you have used it, there is a realization that there is great power that can be wielded without any effort. Till you can try it for yourself see what twitter has to say about NServiceBus. For me the next dose of sharing some NServiceBus love will be at Day 2 of Doing .net Days on the 27th February 2010 if you have nothing on that Saturday why not send an email to register a seat before they run out

That’s ok you say we are used to the Microsoft code name to product name shenanigans. Sure but its way more than just changing names and focus or both. This quite often has me having this conversation with myself

Me: What is Microsoft Distributed Caching product? Velocity?

Myself: No that’s the old name now its “Windows Server AppFabric”

Me: Ok thanks ill download that but what does AppFabric have to do with Caching? And where is the CTP of that application server that came from the “Olso” group? Dublin.

Myself: Oh you have that already?

Me: No all I have is that “Windows Server AppFabric”

Myself: Yes that’s Dublin.

Me: But I thought that “Windows Server AppFabric” is Velocity

Myself: Yes it is

Me: BUT BUT BUT….

Myself: No Buts It just is.

I hope that this constant cycle of change innovation duplication from Microsoft slows down to a pace that people can keep up with. Maybe even have a sniff of the roses and ensure that it is actually making a difference, and not just causing more noise from Redmond.

The first thing that went through our minds was that we had written the persister wrong, however a persistent Simon continued to work through the issues. Along the way we discovered a few things.

1. NHibernate is faster and therefore able to handle the load better.
2. Proper Saga Design is important.

NHibernate is faster

Most of you already know this but the key thing here is that because its faster it fails less therefore the issues that we saw in the development of the persister was difficult to isolate.

In the Manufacturing Sample that comes with NServiceBus start up all the required services and when the partner is ready run it in simulate mode, it does not fail. doing the same thing when using the L2S persister had the same effect.

Once the number of worker threads on the Order Service Host was increased to 4 could use we started to see a whole bunch of errors. From messages not being completed to violation of primary keys and deadlocks. Using the NHibernate persister required us to increase the workers to 6 or 7 you would start to notice similar errors.This leads me to the next item.

Proper Saga Design is Important

If your not sure about what a Saga is please read the NServiceBus wiki. When considering how you build a system  though that goes into how you would like to handle long running transactions is important. Knowing how many transactions per second. What are the SLA’s? My rule of thumb the more i think about it is that a Saga message that starts the saga should be unique any subsequent messages that are part of that saga should be dealt with a more explicit message handler. So you have a CreateOrderMessage that starts a saga and subsequent messages are handled by UpdateOrderMessages for changes to the order, with CancelOrderMessages and AuthorizeOrderMessages being explicit to the saga.

In my opinion doing so produces both a simpler and more explicit design of your long lived transaction. This however is now going down the path of a different post one that warrants a post of it own.

The Benefits of Reading code became clear the deeper we went through the NServiceBus code base, the more you understand how each piece interacts with the other to provide a great framework. Despite all the hurdles L2S provided it was worth the effort. What must be said for NServiceBus is that it is highly extensible already evidenced by its existing OOB support for 4 IOC frameworks, now 3 ORM’s and in the future with version 2.1 it will support a whole raft of Queued Transports.