Biocross Server Management System

The factors in speed are the following:

• Size of your XML Gamedata

• Size of your Player Files

• Power of your CPU (bearing in mind, mine is nothing special)

• Hard drive Speed (again nothing compared to server standards)

• RAM helps too (a higher Read/Write speed for RAM)

please be aware this test was run on my home PC - with tonnes of other processes in the background and not really an ideal server environment…

How long was it on IONCross?

I’ll re-test again on another server and let you know…

2.5 minutes sounds ok for me IFSO needs about 5 minutes for our 856folders with something around 1600char files.
So while expecting that Hazard meaned 612 chars not folders i would say its ok cause it would be a bit faster than ifso.
If its player folders and not char files than WOW thats a good time.

1,220 files are generated (because player data XML and equipment are seperated) there were 327 folders…

that time frame includes the time it took to decode and save the decoded version of all those FL files…

After the generation you will NEVER have to regenerate again - synchronization does all of the following things:

• Checks for FL Files that no longer exist and deletes the corresponding XML

• Checks for FL Files that don’t have an XML reference and creates it

• Checks for FL Files that have a later modified date than when the XML was created then re-creates it (updates)

all 612 files did/do this in under a second…

Attached is some PlayerXML Data for you to look at and examine, also attached is Vanilla FL Gamedata for you to look at:

Let me know if you have any ideas for the format and the layout…

Well, I’m hoping to use it for around 13,000 player files. The time generate the xml conversion doesn’t bother me…30 -50 mins seems fine.

I do have a quick question. Does Biocross deal with both encoded and un-encoded player files. Depending on the flhook settings you can have both plain .ini player files and encoded ones.

Well if they are encoded it will decode them - but when it comes to reading from them it will only do decoded ones unfortunately…

http://code.google.com/feeds/p/biocross/svnchanges/basic - RSS Feed of all the changes made to Biocross as I make them

http://biocross.googlecode.com/svn/trunk/Biocross/ - Read Only version of the code…

FLAdmin - Class Library that makes FLAdmin.DLL easier to use
FLRead - Reads values from decoded FL Files
Biocross Main - Handles all the XML Stuff (synchronization, generation etc.)
Biocross - The main program

Yay for open source

So, would there be a possibility for a light scan or something similar, that like only scans for the character names, so that it doesnt take so long?

Because we have like 50 or 60 000 chars

Not really as nearly every feature relies on that database of players, I will take a look for you see if I can speed it up somehow…

I’ll let you know how I get on

If that massive scan is only needed once, then its not really bothering me anyways

Isn’t it possible to store the data in a DB?
.NET has standard support for that mssql express version. Doesn’t that speed up things? Cos my experience with XML is that it is slow.

Then users would need SQL Express which is quite a large download, also XML is not actually that slow, I personally feel these figures are way off due to how many programs I had open, the power of my PC in general and other limiting factors - I will wait to see what happens when I use it on a real server which should be soon, for now I guess it’s the wait

You could use SQLite, it doesn’t need any services and is only one File, which is opened by the Application itself.
It also IS like a real SQL-Server, but I think it’s slower than XML

I do like SQLLite but I think it is quite slow… either way I’m looking into the speed and refining it etc…

Haha using SQL Server for the same amount of FL Files and same files!…

11 seconds…

yeah thats right 11 seconds 610 files do the math thats 11 / 610 = 0.018032786885245901639344262295082 seconds per file …

Does anyone have any problems me now switching Biocross over to SQL?

Depends how long that pushes the release away ::)

Just like to say college is really slowing me down, but the arrival of now a 4th Server operator has really spurred me on to enter the ‘competition’ as it were, however I guess really this is an oppertunity as looking at all 4 (plasma, disco, biocross and IFSO) they fit neatly into these categories:

Beginners:
IFSO

Intermediate/Advanced:
Disco

Intermediate:
Biocross

Advanced:
Plasma

Anyone got any otherr judgements on that? The thing stopping Biocross being in a beginners category is the requirement for SQL…

Things are turning out well in the server development world…

Hey, do you have a linky to the plasma one. I haven’t been able to find it…and didn’t know it existed either.

On performance matters:

• I almost used SQL but I found that XML backed file storage was fast enough. I just had to be careful on the key selection in use on the tables. I haven’t had a close look at the biocross code but maybe … (fyi I screwed up a key definition and it really (100x times) slowed down record insersion into the tables.)

• The disco source has a C# based flcodec, bini and flhash algorithms in it. I’m sure you could port the algorithms to biocross and then you wouldn’t have to call that dll thingy. (My knowledge is a little fuzzy regarding the dlls for biocross) As I understand it, the algorithms should run at the same speed in VB. I’ve also got a resource string dll reader coming in the next version so it might be useful to you and save you a little work.

I’ll add that it is friendly competition between the programs - and maybe competition is the wrong word although I can’t think of a better one right now.