Looking for a Database Programmer

[tillyoubreakit]

Its pretty funny that some people dont seem to get the "CURRENT" point.

I never said "Thats what we will do"
I never said i wana store 1TB in the beginning!!
I did say a small modular base for scaling later


Quote!!! We figured that we dont have the power right now to run 2 projects at once, so this is delayed till the one right now is done which should be 3-6 month Quote!!!

PS: After this 3-6 month i really dont care if i have to spend 300K on this. But the beginning we will not be even near it ;-)

[scottyman]

Plan for the worst, hope for the best.
It's all well and good designing a small infrastructure that "should" scale - but in terms of the db design, it's better to build an application tested at the scale at which you should hope to eventually run it at.
I've seen some very large, very expensive projects develop cost overruns of more than 5 times the original project budget simply from making assumptions from the outset.
My 2p is all, and I plan for worst case. In all likelyhood you're completely correct, and I'm just flying off into the realm of the "silly".

Primary issues I think you should be looking at, is whether or not it's viable to allow download of video files - how about low bitrate streamed files, with option to auto-burn cd's at the back end, and print shipping labels - that's simple and elegant.

If you're having a GUI driven interface, .NET isn't portable, and I've had some exceptionally simple tools (in comparison to the scale of what you're proposing) written in .NOT fail under the most amusing circumstances... Also, modules written to be cross platform (by the client) have been found not quite be cross-platform... if you meant supporting XP AND 2000, rather than Windows, Linux and Mac that it.
My front end worked, their backend didn't.

Hence, choosing Java, however that has issues of it's own. PHP is nice, you've still got pretty good sandbox functionality, and v.5 brings with it some better OOP constructs, but bizarrely and like Java 1.5, some old fashioned C-style bits too.

DB wise, I would seriously suggest oracle for what you're planning to scale to, but then MySQL has had very good remarks made about it's clustering capabilties. If you're going .NOT, then you'll be better off with MS SQL, as there's less buggering about with the back end (ooo err missus)

Regardless, sounds like fun.

[killgORE]

Access DB ???? Haaaaaaahahah... Just joking. Think its pretty clear that my prefered language is Java. The streams are easy. Jsp/servlets is just plain sexy as a way for developing applications that scale. As long as each class and jsp has well defined, and, a small amount of functionality. Struts is the way I have developed webapps in the past.

The only issue was the speed aspect of java... use java today. Get your results tomorrow ! But at least you get the results properly on most platforms.

[sparky191]

Just to add my 2cents, in terms of searching this data its not that hard. You can build up a query GUI that allows the users to build sql like queries without using sql. Basically it would be an layer of abstraction that would allow the user to use english like language (or what ever language) for sql queries. If you want to see some screen grabs of a similar system I can get some for you.

Hardware wise you'd obviously need two servers, the live and the other a development/test server which would replicate the live server. You can buy tape and optical caddy systems that can back up TB of data, that problem is one of money rather than a technological problem.

How the server would be configured or what db to use I couldn't say. Theres no mention of what sort of query traffic would be hitting this db, so that area of the spec is wide open. Obviously with all the major db theres a lot of tweaking you can do with the hardware and the software to speed up the type of work that they are doing. I've seen large DB's using Oracle on 4 and 8 cpu, Red Hat servers that were impressively fast. We were using ASP and VB (not .Net) and it did all we needed.

As regards portability, I guess it depends if you want to increase your budget by 30-40% to support an extra 5% of users. Personally while I've seen some great and some terrible Java system, I've never thought the portability was worth it the majority of times.