powered by Jive Software

I wanted some advice and maybe have work

I’m planning two setup’s and wanted some input from the community… I wanted some help figuring out how powerful of units i should build as well as what plug-ins to use and how you would configure things like clustering and what type of connection would be needed for such a task.

Project 1:

Needs to be able to handle around 50,000+ users online, used for IM only with no file transfer or video. Must be able to connect to other networks such as facebook, gtalk, aim, yahoo etc. So the end product will be a network that not only allows for its own IM service but allows people to connect to other networks and able to have 50,000+ people using the service in this way at any given time. system must also offer livehelp like fastpath.

  1. How many servers would i need to do this?

  2. How would you build this server (including type of cpu, number of cpu’s, amount of ram, speed of ram, space on hdd ect. be as detailed as possible)

  3. What type of router and/or switch would you use?

  4. How fast of Internet would be good for this?

  5. What OS would you prefer for this setup?

  6. What plug-ins would you use for this type of setup?

  7. how would you setup clustering if any needed?

  8. If i was to build the units and get them online how much would you charge to set it all up and maintain it on a monthly basis?

  9. Would you be interested in becoming a full-time member of our team? and willing to re-locate to NY? if so desired salary?

  10. Any other tips or information you would like to share.

Project 2:

Needs to be able to handle around 10,000,000+ users online, used for IM/Voice/Video (like google hang-outs) with no file transfer. Must be able to connect to other networks such as facebook, gtalk, aim, yahoo etc. So the end product will be a network that not only allows for its own IM service but allows people to connect to other networks and able to have 10,000,000+ people using the service in this way at any given time. system must also offer livehelp like fastpath.

  1. How many servers would i need to do this?

  2. How would you build this server (including type of cpu, number of cpu’s, amount of ram, speed of ram, space on hdd ect. be as detailed as possible)

  3. What type of router and/or switch would you use?

  4. How fast of Internet would be good for this?

  5. What OS would you prefer for this setup?

  6. What plug-ins would you use for this type of setup?

  7. how would you setup clustering if any needed?

  8. If i was to build the units and get them online how much would you charge to set it all up and maintain it on a monthly basis?

  9. Would you be interested in becoming a full-time member of our team? and willing to re-locate to NY? if so desired salary?

  10. Any other tips or information you would like to share.

PS: Budget is moderate to high for this project but lets not build 20 $10k units to get this job done. Try to keep the budget respectable and as low as possible to get this job done correct. The job related question are legitimate as we are likely to need someone on hand to take care of the network, and it would be great if you were skilled in server administration all together not just xmpp. Web development skills also a huge plus. please send resumes to desirehireapp2 [at] gmail . com

-Mugzi

You are asking for an awful lot of information for a discussion board.

You need to do some basic research first, for example, facebook is not federated, therefore your server cannot talk to it. That will have to be accomplished directly from your client. You would also have to investigate whether you can federate with the other servers you have mentioned as well.

You also talk about a moderate to high budget, yet imply that $200,000 is too much and you want to support 10 million plus concurrent users. You will have to define some budgets in actual monetary values for anyone to be able to suggest any reasonable solution, assuming one can be gotten for your budget.

You know honestly, I don’t know why some people find the need to chime in if their not going to add something constructive to the conversation… My questions are related to getting adive to formulate a budget!!! even as i wrote not the build 20 10k units I implied it in a way saying lets not do it if we don’t have to… i left the discussion open to see how other people would do it, to see what others can suggest as a configurations… But somehow like always one person who just… as for facebook I’m not sure if it fed or not but using kraken or spectrum you can connect to them as we’re already doing that now, so being that you didn’t know that maybe this isn’t the right discussion for you!!

Before you go any further, you should first realize he is one of the main (if not THE main) developers of Openfire and Spark… he knows what they can and can not do. What he meant by facebook not being federated is that your server does not talk to facebook, it only allows your clients to pass through and connect to facebook…your sever will never talk directly to theirs. Outside of that, you’re most likely not going to get the answers you want here so don’t be surprised if you don’t. This really is a help and support forum, not too many people talk about what you’re looking for. It’s few and far between.

Have you searched around? These will give you an idea to start with.

The only other possibly helpful thing I can add is that you will probably have the most luck getting an answer through a consultant. Have you tried going through Jive?

http://www.jivesoftware.com/services/technical-consulting

Firstly i wasn’t aiming to offend but rather to instill a sort of rule-set to the thread… If you honestly either can’t answer or wont answer, what’s the point of chiming in?? If the community can’t answer this type of questions then what help can the community be? Especially a main developer… I mean really how can you develop something an not know or rather not want to help when a user asks for advice, and rather spend your time bashing the question itself… At the end of the day, if a software is released without someone being able to figure out how it works, or what setup would be good, or even its limitations then how can someone adapt the software?? I mean I even went further and offered a full time position to someone experienced if they could at least break-down the bare minimum requirements for such a deployment. I work with over 100 web designers and developers everyday and if I walk over ad ask a question of this sort and get the answer I received then that’s the last week that developer works for me. You can’t sell something or rather expect people to use something you yourself can’t break-down.

So in closing if I did offend rcollier it wasn’t my intent but rather a firm way of deterring other answers that didn’t stick to the question at hand. as for you Derek thanks for the information…

-Mugzi

What the question really is… Is if you worked for a company and the above specifications were put on your desk, and you were told all the tools in the shed were at your disposal, how would you achieve the task at hand at the most reasonable cost to the company ranging from building the unit’s serving the projects to picking the right speed of Internet to actual number of units and configuration of the software on the units. There really isn’t a right or wrong answer but rather idea’s to be found on how to make this possible. Maybe if we try we can really make this the end all discussion at the very least to this version of the software so we dont have to read 700k posts asking the same question…

-Mugzi

I have briefly glanced at this post yesterday and just ignored it, and probably we all should do (Robin, Derek) as such complex inquires can’t be answered by this community. Robin is not the main developer, although he does some work on Smack and pubsub support, but there are no full time or main developers here. All projects here were open sourced and left as it is by Jive at some point and now only a few volunteers trying to do something in their spare time. I don’t believe there are someone with so much experience to advice on a budget for a system with 10kk users.

I was going to ignore it as well, but I will just add for the record that this community does indeed do have people that can answer such questions, but expecting a commercial service from us here will not happen.

Any one expecting to buy something should consider commercial solutions from Microsoft, Cisco, etc.

Any one expecting proffesional advice about the open source products here should hire a consultant and pay for their time.

1 Like

Dele Olajide wrote:

Any one expecting proffesional advice about the open source products here should hire a consultant and pay for their time.

Well, he is willing to pay (offering a job), but i don’t think he will find that kind of people here either. Most of such people would already have a job.

I thought I was being constructive.

I told you that you shouldn’t expect much of an answer to your list of questions on this forum (as has been pointed out several times now), pointed out holes in your plan to do server to server with facebook and simply advised that you might want to supply more concrete budgetary information so someone who does have the knowledge you seek could probably tell you up front whether your expectations are realistic.

My apologies if you were insulted by the “do some basic research” comment, it was not my intent, but I can see how it could be taken the wrong way.

Either way, I guess I should have followed my first instinct and ignored the question as well.

1 Like

I rest my case… The answers that followed my most recent post further proves what i was saying and why i even acted the way i did in the first case… As for people of that sort already having job’s, remember there are always better jobs to have, why do you think so many people jump ship from lets say google to amazon and visa versa… from what i gather 10k users shouldn’t be that huge of an issue to achieve as people have said they did 200k on one box using connection managers. I’m just trying to figure out what the capabilities are as well as the limitations. And don’t tell me consultants and such because i would have to be an idiot to waist 100k of my companies money to find out the system isn’t what I’m looking for and scrap it… It would be fiscally irresponsible of me to do that to my company. The honest truth is there is more then one xmpp server to choose from and I’m only looking to make the right choice to fit my needs. It seems in this case so many developers want to chime in with their 2 cents on what they didn’t like about the question, but none can answer the question of what’s possible… I mean really guy’s one said he ignored the question and everyone else should, the other one says get consulting… In my language we have a very old saying “A bad dancers ball’s were in his way”… It’s the easiest response to say ignore the question… It shows how much of a labor of love this software was and is…