Happy new year to those of you for whom the day has changed and of course to everyone else when it happens! I just wanted to cross-post a couple of articles I've written on some of the pet projects I've been working on over this festive holiday season. They're related to the Raspberry Pi, Fuse Fabric, Arjuna/JBossTS/Narayana, vert.x and MongoD. So if you're interested, check out part one and then part two. And once again, I hope everyone has/had a very happy new year!
It's Christmas day and I know many of you will be busy opening presents, eating lunch or just sleeping off the night before! But I did want to take the opportunity to wish everyone a very merry Christmas and best wishes for 2013. It's been a great year for JBoss/Red Hat, what with the release of EAP 6, 3 JUDCons, a couple of acquisitions and a partridge in a pear tree! All of those things (with the one obvious exception!) in conjunction with EAP+OpenShift, our mobile work around AeroGear, a larger than ever JBossWorld/Summit, our continued invasion of JavaOne, hiring some of the key JRuby developers, and much much more go to show that not only do we have our mojo back but that we are breaking down some of those old perceptions that the JBoss brand is only associated with the application server. Our communities have grown, with the addition of new projects such as CapeDwarf, and others have just continued their expansion, like Arquillian. But at the heart of all of this success is one common component: you, our community. I cannot over state how important our communities are to me personally, but also to JBoss and Red Hat. You help to ground us in the realities of what users want now and in the future, shaping middleware not just in the Java space but across many other languages. Whether you provide code to a project, provide feedback, give use cases, or somehow give some other kind of input, you are members of the JBoss family. And for that, I thank you! Here's to 2013!
I've been talking and presenting about cloud and PaaS for a number of years and each time I do, I say pretty much the same thing (and I paraphrase): "I think private (on premise) PaaS will be more important than public PaaS". I won't go into the reasons why, except to encourage anyone interested to check out some of the older posts. That's not to suggest that public Cloud isn't important or useful, which is why Red Hat has been making great strides with our OpenShift offerings, which we're proud to ensure remain open source. Over the last year or more, we've seen Java added to OpenShift, as well as JBossAS and EAP, and many other languages and frameworks are there too. The number of applications running on OpenShift has grown significantly since we officially released it last year and our community has increased much more.
However, although there is still a lot more we want to do on public OpenShift, the feedback we have been getting from community and customers has been the need for an on-premise offering. The intention with any on-premise PaaS has got to be that it's a cloud and not just a rebadged virtualisation offering and I'm really pleased to see that we've stuck to our principles with the announcement of OpenShift Enterprise. All of the things you've heard about OpenShift, such as its mission critical security (based on SELinux), flexibilty (multiple languages, frameworks etc.) and scalability (including, cloud bursting eventually) will be part of OpenShift Enterprise, the first enterprise PaaS. So what about JBoss? Well as I've said several times before, if you want an enterprise PaaS then you need an enterprise infrastructure (middleware), so it shouldn't come as a surprise that your favourite JBoss technologies and products will be coming to OpenShift Enterprise. In fact EAP and EWS (the production versions of the application server and Tomcat) are available already.
This is a very exciting time to be involved in middleware and the Cloud. I'm pleased that we can finally push forward on the larger vision we've been talking about for so long. And what next? Well maybe we'll be able to bring some of that ubiquitous computing cloud into the picture eventually
Devoxx has always been one of my favourite conferences. Back when JavaOne was more product (and Sun) oriented, it was a bastion of independence and what was going on in the wider Java community. Even today, with JavaOne improving, it is still an important conference. And that's even though I prefer it's old name of JavaPolis (what's that they say about trying to please all of the people all of the time?) This year was even bigger for Red Hat/JBoss. Not only did we have our usual healthy mix of sessions on things ranging from Arquillian to Polyglot, but we were a key sponsor. We payed for the party at Noxx, which was a very interesting venue. We also co-sponsored the Hackergarten, which was a great opportunity to meet like minded developers and community members.
Importantly, we also used Devoxx as part of our renaming exercise for JBossAS. However, we did have to change our original plans slightly. Back at JavaOne when we announced the rename, we thought that we'd have started and ended the name collection weeks ago and had a select number of names voted upon by Devoxx. In fact I worried about how we would choose the final number of names to put up for vote from those suggested by the community. How wrong I was. We received over 2000 names and from that there were almost as many unique entries. There were only a few names we obviously couldn't take forward and even less from people who suggested keeping the original name. However, then we ran into the reality of copyright and trademark law! For each name our legal team had to do an exhaustive search to ensure we would not be treading on some other company's name or product. And this search takes time - proportional to the number of names!
What this meant was we simply had no way to get the names ready for vote until just before Devoxx. So we decided to repurpose our Devoxx announcement (at the start of the keynote) to telling people the above story and announcing the names for vote. Now just as some people may not have liked the rename of JavaPolis to Devoxx, Marathon to Snickers, Borland to Inprise, or Coco Pops to Choco Krispies, some also don't like the rename of JBossAS or the selection of names we've ended up with. But from the feedback we've received, including during the Devoxx BOF (more below), it seems that the majority of people so far get it or don't see it as too much of a problem and can work with one or more of the names. Now that we've added more descriptions around the various names (thanks to the AS team expanding on those submitted originally) I hope people can see the potential with them and give it a chance. (If you feel otherwise, then feel free to email me as I value your feedback.)
After the keynote, Ray and I had a session where we tried to give an overview of all things JBoss related. We decided to reuse this years JBossWorld keynote, but without the audience participation in the demo. Ray did a great job of re-running the demo with bots playing the various roles and this was great too, since it shows how much we've simplified the testing and simulation of complex applications. And of course there was the BOF, which really could have run even longer if we'd had the room longer. There was a lot of good discussion and community involvement, as well as beer! And because we're good community members, we let our beer spill over into the adjacent BOFs
In conclusion, Devoxx was a great success. Our sessions went well. The community gave us a lot of great feedback. I got a chance to meet new people and friends. And the general vibrancy around the conference made a fitting conclusion to 4 months of travelling!
So Ray and I just finished our stint at the Devoxx keynote this morning to announce the names for voting. The video should be up soon but in case you don't get a good view of the Dilbert cartoon we used to explain the situation to the audience, here's the link. We were struggling with how best to let people know the difficulties in taking 1800+ submissions and getting them through the legal process, when Dilbert came to our rescue. I don't think we could have put it any more succinctly as this, so why bother?! Anyway, the voting process is now open and we'll be updating the pages on JBoss.org to try to add more flavour/meaning to the names that we eventually managed to get from the great submissions.
Thanks to everyone who submitted a name. There were many great names, but unfortunately we couldn't select the majority of them for reasons mentioned in the above cartoon. But the ones we have are good and I'm sure we can get behind whatever is eventually selected and make it a worthy successor to the JBossAS name! So get voting.
I'm just back from giving two very different presentations in the space of 4 days. The first was at the CME Technology Conference in Chicago, where I was invited to give the keynote. I gave an updated version of our vision of where cloud computing meets ubiquitous computing meets mobile. The room was packed and from the questions during and after, it went down well. The rest of the conference was interesting too, especially as many of the talks added more detail to some of the things I could only hint at within the keynote.
Then it was off to San Francisco and QCon. We were in the industrial track again this year, which had to be expanded to two tracks given the explosion of QCon attendance. Another result of their increased popularity was the move to another hotel, which had the added benefit of allowing everyone a chance to meet much more between sessions. I hope they do something similar for London, since the conference facilities they use there are too much of a rabbit warren.
Anyway, the talk I gave this year was based on the polyglot talk that Bob and I gave at JBossWorld. There were about 30 or so people in the room, which is good given the other tracks they had at the same time. And importantly, the audience asked a lot of questions throughout the presentation on our strategy, community interactions, and JBossEverywhere. I hope that some of them were interested enough to get involved with what we are doing and perhaps even start some similar projects in languages we can't cover at the moment. QCon takes immediate feedback from the attendees and then sends it to the presenter; the results of that were overwhelmingly positive too, so I'm hopeful that the projects I mentioned will see some benefits.
It's Devoxx in the next few days and raft of presentations and tutorials by JBoss people. I've got a keynote to give and maybe more. But once it's over that will mark the end of almost 4 months solid travel and presentations. Well almost ... There's one more to give in Boston in December, but then I'll be close to Christmas vacation and a chance to rest as well as catch up on some of my personal projects that have not had the attention they deserve recently, such as STM, mobile and Erlang.
I just wanted to post a quick status update on the renaming of the core JBoss Application Server project. As you all know, we asked for the community to send in their suggestions for a new name a few weeks back. We were overwhelmed with the number of names we received (well over 1500!) With the benefit of hindsight this shouldn't have been a surprise, but the amount of names we received were a lot more than we expected. Unfortunately this has had an impact on our timescales. Whereas originally we thought that we'd be putting up the more popular names for a community vote by now, that hasn't been possible because we have had to go through these names and check them against trademark infringements etc. This takes a lot of time, proportional to the number of names we received and we're only just drawing that phase to a conclusion. So what this means is that rather than announce the results of the vote at Devoxx next week, we'll be announcing the names that we're going to vote on. At this stage we're looking at a voting period of a few weeks and that then takes us into the Christmas holidays, so we'll probably make the final announcement early in 2013.
I'm going to be giving the keynote at the Red Hat Developer Day in London in a couple of days time. You should come along and hear what we have to offer to developers and also what we'd like developers to help us with. I'll try and hang around as long as possible to speak to people, but unfortunately I'm actually on holiday this week and it's my wife's birthday the day of the event! So apologies in advance if I miss you this time around.
I'm just back from speaking at CLASS (Cloud Assisted Services) Conference in Bled, Slovenia, where I was asked to speak on the Red Hat vision on Cloud. Ales was also there talking about CapeDwarf and Trustin gave a tutorial on building applications on top of Infinispan. There were a lot of other speakers talking about their experiences and vision for Cloud, and it was good to meet some of them and see that our thoughts are in line with those of others. Unfortunately I was only able to spend a couple of days in the area and just one day at the conference, but it was still worth attending. Some areas of interest included listening to Ken Ducatel speaking about what the EU have planned in terms of security and authentication standards for all cloud vendors operating wthin the area, and then real world Cloud uses from Kate at Argonne National Laboratories and Christine from AmpLab.
But another important reason for attending the conference was that it was my first time in Slovenia and to meet the entire team based in the area. I was already impressed with the team (hey, they're JBoss engineers, so what else should you expect ?), but Slovenia was much more beautiful that I had expected. Upon reflection, if I'd done my homework or been awake on the flight from Paris to Ljubljana airport, I'd have seen much of this from the air. However, I didn't and it was a very pleasant surprise. I hope to have further opportunities to visit the area in the future and next time I'll spend more than a couple of days!
I'm just back from JavaOne, where once again we had a great turn out for JBoss related sessions (almost a dozen!) I'm sure the other speakers will post blogs, if they haven't already, to give everyone an idea of how the event and their sessions went. My own were well attended and judging by the various comments in the talks and afterwards seemed to appeal (slides should be made available on the JavaOne site soon). I know that the "meet the EE7 experts" session was also well attended, but this time I was in the audience and Pete Muir did a great job of explaining some of the things we'd like to see improved in the standards as well as the way in which they're managed.
Of course one of the things that has quickly become a tradition with JBoss (and a success) is our booth and the booth sessions/presentations that we do. There were 21 sessions this year and every time I swung past the booth it was packed! Many of them were recorded so they'll be online eventually, but this picture I took gives you an idea of the level of interest.
And of course another tradition is the JBoss Party. We held it this year at the MOMA and although it was slow to get going, it quickly gained the momentum that is typically associated with out parties!
Overall the feeling throughout JavaOne was great and I think JBoss and Red Hat participation was key to this. So here's looking forward to next year!
I had thought that we had been pretty explicit about the reasons behind the recent announcement of the JBossAS rename, but some feedback I've received makes me think that another entry is necessary. As I said before, there are a number of factors involved in the rename decision, some technical and some not so. But one of the major issues we want to fix is the confusion inherent in the term 'JBoss'. Sometimes people use it to mean the community. Sometimes it's the middleware group in Red Hat. Sometimes it's the EAP product. And sometimes it's the community application server project. This confusion can cause major problems, especially if you download JBossAS (AS7) assuming you're going to get a fully supported binary, only to be told weeks, months or years later that that isn't the case!
So we find ourselves at a point where, because the project and product often diverge over the lifetimes of versions, it creates a lot of confusion when they have the same name. Instead, when people see software called "JBoss," we want them to know that it's a commercial product with a well defined SLA and support lifecycle, in contrast to the community projects with best effort forum-based support provided by the community. And other projects with "JBoss" in the name will continue to be renamed, as we've already been doing for the past several years with messaging, transactions, security, etc. We also hope that more clearly distinguishing the projects from the products in this way will encourage wider community participation.
When you think of major influences on Java middleware over the years then JBoss the company and JBossAS the project should be in your list. Whether it's bringing J2EE to the masses, helping open up the Java Community Process, or simplifying application development, JBoss has played a significant role. However, as I've pointed out several times, the middleware landscape of today has changed and is continuing to change radically and we continue to change with it - in fact I assert that we are driving a lot of this change. Our plan, best summarised around JBossEverywhere, is obviously to continue to drive for further improvements in the EE standard and specifically in our implementations, e.g., DeltaSpike. The JBossWay initiative is the project under which many of these changes will happen. Of course we have far more today than just an EE application server, so JBossWay will extend to those as well.
Our work with mobile and embedded devices, such as running projects including AS7 on a RaspberryPi, helps to put the capabilities that all enterprise applications need into the hands of developers no matter where they are. Of course the various polyglot work such as with TorqueBox and Ceylon, is equally important because although we believe the JVM is the right base on which to build, we don't want to force people to use Java. For the kinds of cloud PaaS that we are focussing on with OpenShift, this flexibility is critical. It's also flattering to see others trying to copy us! And talking about cloud, CapeDwarf leverages JBoss EAP6 underneath GAE interfaces in OpenShift, so JBossEverwhere truly goes from the small scale to large scale!
As I mentioned before, the JBoss name now covers far more than an EE application server. We have SOA, BRMS, Portal as well as tooling, and with the acquisition of FuseSource and Polymita our breadth and depth of technologies increases. JBoss morphs every time an acquisition occurs; in fact every time we hire someone we evolve, being influenced by them and influencing them. So really when I talk about JBossEverywhere I'm talking about JBoss/Arjuna/Metamatrix/Fuse/Polymita/.../Everywhere. Mobile and cloud have driven another area of innovation: NoSQL and BigData. With Infinispan and JDG we announced part of the strategy there. But much more is to come in the near future. I can't go into details just yet, but suffice it to say that I've been working with teams across JBoss and wider Red Hat to formulate a plan that will take shape over the coming months.
JBossEverywhere is intended to put JBoss technologies into the hands of as many developers as possible. The recent JBossAS project rename announcement and associated strategy to enable developers to gain greater access and availability of our products, is therefore yet another "string in our bow". I am sure that by allowing all developers to use our products for free whilst at the same time ensuring innovation in our communities, is not only a game changer for professional open source, but it also helps all of our communities grow and benefit from each other. Of course there's a long road ahead of us with all of the plans I've already mentioned, as well as those that I haven't. JBoss will continue to evolve. JBossEverywhere will influence much of what we do in the coming months. And I'm hoping that with the changes we've announced recently that all of our communities will help make it a success.
I first came across JBoss in 2002 when started Arjuna Technologies and we were looking around for ways to leverage the transactions and messaging pieces we'd brought with us when we exited HP. We subsequently replaced the JBoss transaction and messaging pieces with our own and had a fairly successful business selling that combination. Fast forward a few years and JBossAS was dominating the open source J2EE landscape. With the evolution of enterprise Java, EE5 and then EE6, the Red Hat acquisition of JBoss, and a few other changes in our industry, JBossAS evolved and became the basis of everything we do. Whether it's portal, SOA or mobile, JBossAS is the fabric on which we build. And then of course there's the Fabric effort which is continuing to feed into it and other projects.
What this means is that JBossAS today is very different to what it was a decade ago. The uses cases it has to satisfy are orders of magnitude more complex and diverse than they were. Put simply, although Java EE is central to what it does and we do as a community, it is no longer the only driving force behind it. And over all of these years, through all of the changes I've mentioned, we've all known it as the JBoss Application Server, or JBossAS. With EE6 came our shorthand of AS7, but it was really still JBossAS. Well now we'd like to change that. We've had lots of discussions here and we believe that the time has come to change the name of our project to better reflect the changes we've seen in its reason for existence so far, but also for what's to come in the future. Importantly with something so influential to open source, we want our community to be involved in the rename in just the way they have in other aspects of its life. So this is your chance to suggest a name that will be associated with the number one application server.
I'm just back from StrangeLoop and the first word that springs to mind to describe it is ... Strange. It's definitely a unique conference, covering a breadth of topics that used to be the mainstay of events like this two decades ago, before we had the more targeted efforts we see today. In some ways it worked really well: I sat in many sessions and learned a lot about things I probably wouldn't have known about any other way. But in other ways it felt confused and rushed. Some of the talk, mine included, were pushed to just 25 minutes, which is not enough time for most topics. And then there was little to pull the individual sessions together into a coherent message or theme. There were tracks, but they were often split across many different rooms requiring people to move around a lot; this may be a good thing to encourage mingling, but it can also be confusing if the "tracks" have sessions in parallel. At the end of the first day I knew I'd been in many interesting sessions, but most of the time I wanted to hear more and couldn't. As someone with me mentioned, it was like eating tapas!
Now don't get me wrong, I thoroughly enjoyed my time at StrangeLoop and come away with ideas I may not have had otherwise. But I can't help feeling that I was either missing something or that the conference was. However, the venue in the Peabody Opera House was great!
But after three weeks of traveling back and forth across the Atlantic, my last image summarised how I felt sitting in the airport on the way home.
Well in case you haven't seen the recent announcement, I'd like to welcome Polymita to the rapidly expanding JBoss family! This is yet another acquisition in the SOA space and I think it compliments our Fusesource acquisition as well as the work we've been doing around jBPM, BRMS and Drools. This is a really exciting time for JBoss/Red Hat and I'm looking forward to working with the Barcelona team. Congratulations and onward!