It has been almost 3 weeks since I began to know windows media services. As I dave into this space, It seems that a dead corner appears when considering authentication plug-in. I looked through the media service SDK attentively but found no clue to this problem. Over the last few days, I have frequently accessed the related newsgroups and forums to try to find its corresponding solution. No luck, again. This morning, I opened my mailbox and checked mail as usual. How I was looking forward to receiving several useful discussions!

 I described my problem briefly and post it to the newgroup. Since I have to wait 12 hours for the response due to the time zone difference, I began to read others' threads with the intention of whether I could find minimum valuable clue about my problem. Finally, I found that some guy was in the same situation as I did. I quickly replied to him and asked him for his solution, but due to the aforementioned reason, I have to wait until tomorrow morning for the feedback.

 Another one said himself that he had implemented the custom plug-in, but can't enable it in media service. He post the related error messages within which the server complained that some dependencies of some assembly missing. I can't remember the error code HRESULT clearly. So I want to reproduce this error on my own system, but another thing happened. I followed the instruction of the SDK, but when I "Add References" to project, the windowsmediaserverice.dll did not appear in the list while the SDK did said this. I had to browse to the folder and selecte manually. But another thing! After I adding necessary import directives, the VS intelligence list functionality did not work. hehe~, so a lot of trouble!

 these trivial problems did took me much time. But it is these problems that made me attain experience. Fuslogvw.exe, the assembly loading log tool provided by framework SDK which I have never used proved that it is helpful when debugging or locating such assembly binding problems as

FileNotFoundException, BadImageException and so on. At least, I had no such experience on this a year ago.

1. The high level perspective of the steaming media system
2. Cache/Proxy mechanism
3. Media conent security strategy
4. Performance improvement (load balancing and clustering)

Typically after you have installed the windows media service, setup a streaming media system is a steps process.

1. Authoring media content. Media service has already provided several tool to help you finish this task.
2. Save content on media server and assign it to a publishing point. The publishing point can redirect request for media content from user to media server. There are two type of publishing point, one is on-demand and the other is broadcast. Simply put, if you want the user to control playback of the media, use on-demand publishing point. Otherwise, use broadcast publishing point. As to the on-demand publishing point, you can choose unibroadcast or mulitibroadcast, but you have to rely on multibroadcast if it is a broadcast publishing point. The difference between unibroadcast and mlitibroadcast is that mulitibroradcast can accomordate more concurrent connections so as to improve bandwidth efficience.
3. Wrap the publishing point as a web page on a web server. You can use the provided wizard style to complete this task.
4. User click the link to fire the publishing point which connects to media server and make media server to stream content to user over internet network.

That is all what I have got today. I will keep going on tomorrow~

MS stated that Windows Media Service had provide perfect solution for Video On Demand situation. Considering our current requirements (a bit like online movie cinema), the most important is that we must install this service on Windows server 2003 prior to delve into it further more. Only do this can we know what functionality it can provider, how far it is away from our requirements. So next week, I will be focus on this task. And another thing I must do is to resume to study the Training cource "How to analysis requirements and define MS solution architects". This course has been divided into several modules. As to me, I would like to study the "Concept Design" module and the "analysis requirements" module. In reality, I always lay me in a situation: so many information around within which only very little I need. I hope to get some instructions and tips about how to do with this situation, what's the criteria of the granulity when dividing related information?

The other day, I saw this piece of news: SilverLight V1.0 had been released. Today I saw that V1.1 beta released news on MS site. So quickly!

This afternoon, I downloaded V1.0 to experiment it. I was told that VSTS SP1 should be installed so that can some VS templates be installed within VS IDE. After long wait and pacience, I downloaded SP1 but failed to installed it cause of unexpected error! Frustrated! Go on, tommorrow!

C# enables developers to use #if precompile directives to state that if some symbol has been defined, then some statement(s) will be executed. It is a common scenario that we always want to do some test by writing some statements when developing program (usually debug mode) but want to remove it after deploying. That's where #if precompile directive can take effect. if we write the following code snippets:

#define DEBUG // this line must be at the most top of the source code file

...

#if DEBUG

#endif

...

 

the red highlighted code will excute or not in term of whether the DEBUG symbol has been defined.

 

Assertion enables developers to state that if some condition is satisfied, the execute flow will continue. Otherwise it will pop a window to ask u to debug or retry or ignore it. In this way, we can make sure that some condition must be satified before some given operation. Both the Debug class and the Trace class have an Assert() method to test some condition accepting a boolean value as in parameter, the only difference is that Debug.Assert() overload methods won't evaluated in Release mode resulting in no performance degrade, but Trace.Assert() will be evaluated in both Debug mode and Release mode resulting in some minor performance overhead. Assume the folowing codes:

 

...

int i=0;

...

Debug.Assert(i==5);

Console.WriteLine("i detected to equal to 5");

...

 

if we must assure that some conditon must be satisfied before some operation, we can use assertion to meet the need.