I’m happy to announce that we’re making most of Openfire Enterprise Open Source! First, a bit of context: for the past couple of years, one way that we (Jive Software) have monetized our Open Source work on Openfire and the other projects on igniterealtime.org has been through Openfire Enterprise. Openfire Enterprise addresses the Enterprise Instant Messaging (EIM) market by adding rich reporting, archiving, and control features on top of Openfire. Since we released Clearspace last year, Jive has become super-focused on social collaboration and communities. That’s pretty different than the EIM market and it’s become increasingly difficult for us to serve both markets with our limited resources. Instead, we want to focus our Openfire work on real-time social and collaborative features and monetize our Open Source efforts through Clearspace integrations.
Discontinuing a commercial product is always a difficult decision and one of our biggest concerns is not leaving existing customers in a lurch. We’ll continue to provide support for Openfire Enterprise through existing support contracts and believe that making the Enterprise components Open Source is the best possible outcome for customers given the options. We remain strongly committed to the Openfire project and are pretty excited about what’s coming in the future.
A Few Details
Gato will have a follow-up blog post with a lot more details about what we’re releasing as Open Source and how, but I wanted to highlight two items. Sparkweb is our flex-based web client based on XIFF and will become Open Source. The client is already very feature rich and polished, and we’re actively making many code improvements to it, as it’s a shared code base with the real-time client features we’re building into Clearspace. Second, the clustering functionality in Enterprise will not be made Open Source. Part of the reason for this is that we use a third-party commercial library for clustering that can’t be Open-sourced.
Let’s Go Get 'em
One of our hopes with this move is that the last possible objection to deploying XMPP-based instant messaging at every organization in the world is now removed. Now, everyone will have access to an open standards solution that satisfies all the needs of IT departments… for free. We think that’s great news for the community and getting our technology deployed even more widely is good for Jive Software as well. We hope you’ll join us in spreading the word.
My only fear is that openfire will now achieve the same status as spark when it comes to upgrades… virtually non-existant. I have said this with a joking tone in the past but now I am much more concerned, that your focus will shift further away from maintaining and improving spark, and openfire. Since they are not paid products what real incentive is there. Although I love your products, I have been very discouraged by the fact that Spark has been in limbo since summer 2007 and now the same could happen to the server.
Let me put it this way. Communication is required for good collaboration. Openfire will always be aligned with our real time collaboration solutions in Clearspace. That means that Openfire is here to stay.
In the particular case of Spark the problem was that Spark’s architectire was not going to meet our needs. We wanted to have the same solution for desktop as for web pages. Moreover, we were looking also for good audio and video solutions and Java was not the best for that. The good news is that SparkWeb it is now open sourced and officially the replacement of Spark.
Now this comment makes me even more concerned. We use Spark because it does SSO as well as the other features (screengrab, file transfer, etc). This was the major selling point of the Openfire/Spark platform in this company. A web solution is not ideal by ant stretch. It has a place for the road warrior in the company, but in general we rely in the ease of spark and its integration into our windows active directory domain. The abondoning of this client/server/domain relationship will not bode well with the powers that be.
Spark is not being removed from the list of products offered by igniterealtime.org. Many months ago we mentioned that our focus was SparkWeb and Spark was in maintenance mode. Those facts did not change. My intention was to say that SparkWeb is the client that we are now pushing and since it is open source then we can consider it the replacement of Spark.
What will happen to the Spark Skinning Feature? It was free or almost free some time ago and then available to Enterprise users only. When you now open the Enterprise version to open source, will the skinning server stay online. I think we really want to know what your plans are on the desktop. Web is nice, Web2.0 is even nicer but imho nothing beats a real app running on my local machine. Is there any chance of getting an *.exe wrap around the Sparkweb_flash_applet???
While I haven’t done anything serious with it yet, I had a quick proof of concept SparkAir running a couple of months ago. Basically it takes SparkWeb’s base code and brings it to the desktop. A number of things have been done to the SparkWeb source code since then to make this even more of a possibility. Interestingly enough, the Air architecture seems to provide easier and in some cases better integration with the desktop than the Java APIs would. I am hoping to have some time to play with this more in the near future and get something whipped up for folk to play with/try out sometime soonish. But I can say I was very impressed and in fact surprised at how well it ended up working. I no longer feel like we can’t do with SparkAir what we could do with SparkJava.
Is this already available? I look forward to trying out the web client and fastpath. I share the concerns/frustrations of mtstravel. It seems that the spark client is very lacking for enterprise clients. Maybe the web client will solve my deployment problems, but man, a spark lite client with only absolutely necessary functionality, with a group policy deployable msi out of the box would be fantastic. When will the enterprise plugin be available?
Daniel, David, and Gato, you’ve mentioned that SparkAIR won’t be able to do Fastpath. Fastpath is one of the main reasons we went with Openfire/Spark for Enterprise IM, and has gotten us a lot of kudos. What’s going to happen to Fastpath?
Getting SSO working via web browser is a challenge, but is very possible. In fact, programming the client side is easier since the browsers already have some support for it. The big trick will be on the server side. I have already started experimenting with PKI-based SSO with sparkweb and Openfire, and I think I can make that happen fairly quickly. Kerberos based SSO may be a little more tricky. Ive done kerberos auth on Linux with Firefox, and I know IE has the ability as well, but that was working with Apache where most of the functionality is already done. Openfire will take a little work get that going for SparkWeb. Its certainly not impossible, though.
But in the end, Im with you. Spark needs an overhaul, and a web-based solution will not be able to 100% replace a desktop app.
I agree…a spark 2.6 or 3.0 would be welcomed. I’ve posted some of our issues before with spark and it just seems like we have to find workarounds. More and more it is looking like we’ll have to have a custom built client made for us. a spark-lite client if you will…Maybe the client control plugin will help. but there are just too many funny quirks with spark.