Service with wide API organizational & techincal questions

Question

Here are some task details:

1. The key functionality of service is Java socket-server (based on Red5. I know it might be overkill, but it makes me confident in future) on unix-platform (FreeBSD or Linux Debian, but I suppose, that the last one).
2. Service deals with quite a lot of simple data.
3. Service must have very strong and wise API. We are also planning to implement something like FQL, FBML, etc. API should be powerful enough to make variety of different sub-services, which should be available even outside of service. Also this API should have different levels of access - for example, some applications are certified only to do some very basic things, while others can even make INSERT-queries in their table (this is just an example).
4. It is possible, that we will need some load-balancing.

Questions:

1. How many developers (and what skills are especially important for them?) are needed for this task to perform (appoximately)?
2. How long (minimally) will it take to develop this kind of application with the given up command?
3. Are there any pieces of advice on how to manage and organize this kind command, if financial motivation is not very big?

And one of the most important question:

1. What utilities and practices may you advice for that kind of project? What is the best infrastructure for it? How all this is usually done?

Thank you very much for your answers!

Answer 1

I'm not sure what you're trying to achieve here so let me try and summarise based on what I read between the lines, and give a suggestion:

• Complex application (strong and wise API)
• Complex environment (load balancing / security model)
• Sounds like the team doesn't exist yet - still planning size of team.
• Unknown scope
• Inexperienced management
• Lack of financial commitment

I would suggest you simply don't have enough information to plan this project yet. If your company is serious about this they need to commit some funding for at least a business analyst and maybe an architect-level engineer for a short contract to assess the feasibility of what you want to achieve. The outcome of this will give you two things:

• The ability to make an informed decision on whether to start the project
• Better information on which to base an initial scope and estimate

Answer 2

I totally agree with you.

But the problem is that our boss is not from IT, and he doesn't understand some difficulties (he had a few Pascal lessons at school and thinks, that programming is very easy).

I've managed to ask for 2 Java-developers, 2 flashers, 2 web-designers. I am sure it's not to much for that kind of projects, but it's better than nothing.

Our boss'll never agree to recruit someone like software architect even for some time. He thinks we're able to deal with this on our own.

That is why I ask for some architectural and conceptual practicies for developing that kind of APIs.