jump to navigation

JBoss Entering the Telco Market June 21, 2007

Posted by Sacha in JBoss, Moved from JBoss.org.
add a comment

We have been working on it for quite a bit of time and now it is a reality: JBoss is entering the Telco market with a unique offering.

The Telco market is currently going through a major paradigm shift. Let me go through some general background information on that shift.

Telco 1.0

The good old telephone network is usually based on a (mostly)voice-dedicated network featured as the “intelligent network”. What it means is that all of the services/features are provided by the network itself while the devices hooking into that network are so-called “dumb devices”. While an “intelligent network” might seem like a great idea at first, it is not really. First of all, it means this network is pretty much only used for voice-traffic: that’s not a good way to amortize it. Second, these networks are usually based on proprietary technologies which makes them very costly and almost always a show-stopper when you want to migrate to another vendor. Last but not least, because the “services” are hosted by this “intelligent network”, the deployment of any new service/feature requires an update of significant parts of the network: this is slow and costly and doesn’t allow user-specific services. Furthermore, if you want to make that new service (let’s say, video) available between two “very remote” users, all operators sitting between these two users must upgrade their network to provide that services, what’s more, in a compatible way.

Telco 2.0

So, where is the Telco market going? It is moving (this started quite a while ago in fact) towards the “dumb network”. While calling your future strategy a “dumb network” might not seem a very smart career move at first, well, it is. Leveraging a dumb network means that the next-gen Telco will be able to leverage the omnipresent and highly redundant IP network called the Internet (possibly backed by private redundant paths) for all of their “communications” (voice, messenger, video, smoke signals, etc.) and services. That provides two key advantages: first, it is cheap as there is no need to build and maintain a dedicated network and, second, it is very agile (Need to setup a new route? Want to upgrade to a new IP router vendor? Want more bandwidth? Easy.) Then, as you know, IP networks do not own any “intelligence” outside of the base routing features. Which, in turn, means that the features are hosted in the devices. That is what makes this network dumb; it hosts no logic or services. The devices do: they become “intelligent devices” – see the twist?. What that means is that if you want to deploy a new service, it is enough for the two end-users to upgrade their software or hardware equipment and voilà, they’re done: there is no need for the core network to be upgraded. Obviously, that’s a simplification, but you get the idea: the Next-Gen Telco (Telco 2.0) will run on much cheaper and flexible networks and provide new services that would remain a nice dream with the good old legacy Telco paradigm.

As with any paradigm shift, existing players have to migrate to a new territory while protecting (and growing) their existing revenues. While the IP/data market is growing faster than the voice-only market, it remains 8 to 10 times smaller today. One of the consequences of this is that operators will have to innovate and create new services that can compensate for that transitional gap. At the end of the day, the great winners in that game are the customers: basic service will be cheaper and they will be able to benefit from (and pay for) new innovative services.

Consequently, Telco 2.0 is as much about a new agile infrastructure as it is a challenge for existing operators, an opportunity for new ones and a new landscape for Telco vendors. OK, I am done with the educational part. Amazon is full of great books on that topic if you are interested to know more about that passionate topic.

JBoss, where do you want to go today?

So where does JBoss and Red Hat fit in this picture? We announced this week that JBoss will lead the development and productization of the Mobicents project into a fully supported “JBoss Communications Platform”. Mobicents is the first and only Open Source Platform certified for both SIP and JSLEE compliance, complementing J2EE to enable convergence of voice, video, instant messaging and data in next generation intelligent applications. Ivelin Ivanov, founder and lead of the Mobicents project and a long-timer JBossian will lead our Telco development and strategy. If you are interested to contribute to Mobicents, here is a good way to start.

The availability of a strong JBoss-led solution will further enable and accelerate that Telco 2.0 paradigm shift. It also means that Telco vendors and operators can now get a complete Telco 2.0 stack, covering both the OS and the middleware layers, from a single OSS vendor. Telcos using our platform will enable the community of creative open source developers to contribute higher level buildings blocks, thus nurturing the ecosystem of new converged applications. That is unique. Unique.


More information is available on Ivelin’s blog



JavaOne 2007 June 15, 2007

Posted by Sacha in JBoss, Moved from JBoss.org.
add a comment

I know, I certainly look a bit “has-been” to only blog about JavaOne now, but guess what, I had the choice between i) being one of the 2’000 guys blogging about JavaOne *during* JavaOne or ii) being the only one blogging about it one month after. I’ve opted for differentiation on that one.

I was positively surprised by JavaOne this year. Attendance was good, show-floor was busy, some very good technical sessions and, more importantly, the spirit was high. What’s more, the JBoss booth was very busy, and not only because we gave away very nice T-Shirts (congrats to the JBoss.org team!). Many users showed up at the booth and spoke about their existing and future projects and migrations. While it is always difficult to estimate the proportion of JBoss users vs. customers and what that proportion will be tomorrow, there is clearly much much more to come. We are just at the beginning of this massive OSS-led market disruption. Much like for global warming, the sea level of software market consolidation is slowly but surely rising.

There was also loads of interest in EJB3/EE5 and how to migrate legacy applications (either based on EE1.4 or Spring) to EE5. Many companies that thought EE1.4 was too complex opted for some non-standard “helper” tools/frameworks/wrappers to achieve their IT goal more efficiently. However, I haven’t met a single user or customer who wanted to keep this situation in the mid- or long-run. That’s good. It means the market is increasingly and inevitably attracted by Open Standards. Today, companies are more conscious than ever of the cost of locking themselves into a non-standardized API/product, i.e., disastrous. We call that “the cost of exit”. When companies perform TCO studies, they take into account a large number of parameters, including migration cost, team education, new licenses, support, etc. but they very rarely take this “cost of exit” into account. Big mistake. In many cases, this cost might be prohibitive and put a halt to new projects or to integrations with new systems, which means you’ll have to live with your locked-in solution until you decide to drop it altogether, which might not be possible before a very looooong time.

On the JBoss project front, Seam and Web Beans (JSR-299) generated a lot of interest, sessions were very busy and people loved its simplicity (integration with Google Guice was also very positively perceived – bye bye XML configuration hell). In fact, the Seam book ended-up being a JavaOne best seller AND the “best stolen” book at JavaOne. BTW, Seam integration with Web Services and JBoss ESB are on their way, so stay tuned.

Also a special mention to “Hibernate Search”. The problem with this project is its name – it is too “shy”. When you hear about “Hibernate Search” for the first time you probably think “well, yeah, it is just a way to populate an index file based on you Hibernate CRUD actions…no big deal.” What a mistake! When you actually look at what it does, how flexible it is, how simple it is to configure it and how smoothly it integrates with the EJB3/Seam programming model, you’ll quickly realize how powerful it really is. After all, “Hibernate Search” is in fact a pretty good name, it is just just too “simple” (I am convinced “Hibernate FreeParisHilton” would have generated more interest). Go Emmanuel, go!

Last but not least, the Open Sourcing of Java is some kind of recurring party. We partied when SUN announced they’d do it, we partied when they actually OSSed the JVM, then they OSSed the JDK minus some classes, etc. SUNW’s PR team must have never been so busy than in the last 12 months. Anyway, the reason I am so excited about all of this is that what SUNW has built is a *very impressive* multi-billion dollar ecosystem that has gathered almost all top software vendors and this, for more than 10 years! But all ecosystems have cycles and >10 years is a long time. Something big had to take place to rejuvenate it and this is exactly what just happened. Now that a quality JVM+JDK is available in OSS, I predict you’ll start seeing a new batch of innovations in the Java space. In fact, it already started, talk to Bela Ban if you want to share some ideas on the standardization of byte-code enhancement at the JVM-level. We just started a new 10-year cycle, so tighten your seat belt.