perfsonar development work

[Maciej Glowiak <>]

Jason,

Thanks for your quick reply.

Jason Zurawski wrote:
> Maciej;
>
> > What's the status of multi-LS? Is there working version on SVN (in the 
> > branch) now?
>
> I am helping to speed up our other worker (Yi) with his BDB work.  The 
> current branch should be good enough to download and install though 
> (contains the necessary libs in the repository, has ant configuration 
> information).  There are changes I am holding back until Yi finishes a 
> little bit more development.  His deadline is very soon, I will push to 
> get the changes in faster since we are holding you up (sorry about that).

I am still curious what BDB improves? Well, I am quite sure it may be 
faster than eXist, but

  - were there made any performance tests and comparisons between eXist
    and BDB? I am sure you did something like this, but I didn't see any
    results and conclusions

  - how far are both DB compatible? I thought BDB didn't support
    XUpdate, so single LS functionality wouldn't work with BDB unless we
    changed all database modification instructions. But for what? XQuery
    Update extension?

  - but does BDB support XQuery Update extension?

  - is XQuery processed the same by eXist and BDB?


> > I am not sure if I can go on with my changes of multiple XML files in 
> > the collection.
>
> Thats a shame, it looked to be promising for performance.

I think I'll go on work on it next week. So that I wanted to have a 
separation between mLS functionality and sLS classes.

> > I also asked you whether it's possible to separate mLS changes from 
> > singleLS? I saw you modified LSRegisterAction. Would it be possible to 
> > divide the main method on some smaller and move mLS code to other 
> > methods? I could help with it, but first I must know if the changes 
> > are final. So, what is the time schedule for integration mLS with LS?
>
> This is up for debate, in the beginning I was under the impression the 
> code bases would be separate (and it is still not too late to do that) 
> and because I didn't make it to the last meeting I am unsure what the 
> consensus is regarding keeping a single LS or going for a full mLS.  

I think LS (sLS+mLS) should be regarding as one LS, but I'd like to have 
code separation between multiLS and core LS. In fact multiLS should be 
independent protocol from LS. And in fact it is.

> Anything that is in SVN (except for some of the code cleanup, extra BDB 
> stuff and some better documentation) is final.

OK, that's important information.


> > Other my concern was on sending request and waiting for the response 
> > in request processing action (as far as I remember it was sending a 
> > request by LSRegisterRequest action and waiting for response).
>
> I was using timeouts to speed up the process of waiting for responses.  
> We never really get a 'hang' situation that way, but depending on 
> latency and other network issues this could be a problem.

You probably tested it in lab. In real 'bigger' network we may 
experience strange behaviour... :)

> > What are berkeley DB support changes for? Are we intend to use 
> > berkeley DB for LS or just for multi-LS? I am not sure. I thought mLS 
> > would use just exist DB. Or these are just separate changes done by Yi?
>
> We intend to install 2 mLSs here at UD and evaluate each database.  Yi 
> is currently seeing some good performance from BDB which is why we tied 
> his project to mine in an effort to offer people choices when installing 
> the mLS.

OK, I fully agree, but my previous question are still open.
Next thing is if BDB will follow "multiple files in the collection" 
changes to LS?

Maciej


--

 --------------------------------------------------------------------
| Maciej Glowiak        Network Research and Development            ||
| 

     Poznan Supercomputing and Networking Center ||
| (+48 61) 858 2024     http://monstera.man.poznan.pl/              ||
 ====================================================================