BEA: The Growing Stable/A Hard Act To Swallow (notes from BEA World 2006 in Prague)

 

BEA’s management team deserves a lot of credit for the company’s performance during the industry downturn which began in 2001 and apparently ended this year. The company isn’t just financially stable, but is now on the trail to becoming a two billion dollar revenue company.

Its not so long ago the first question for CEO Alfred Chuang was inevitably “can you remain independent”? I certainly had my doubts.

But BEA would be a harder act to swallow now. When we look to landmarks in ISV growth the billion dollar revenue mark is the most significant milestone. Getting there is incredibly hard. There are only a few companies that have managed it – companies like Adobe, Business Objects, CA, Google, IBM, Intuit, Microsoft, Oracle, SAP, SAS, Sage, Symantec.

BEA passed $1bn some time ago, but last year it made a company-changing acquisition to punch through that natural barrier into the next phase of growth, in the shape of its Plumtree purchase, which has certainly helped new license sales as BEA pushes more aggressively into portals. A fair amount of its current growth however is organic, based on its Aqualogic service infrastructure line. BEA’s current financial solidity is based on diversification: growing its stable of products.  You could say it has already made the “difficult second album”.

So what’s next? For album number three it appears BEA is going to blog and tackle into collaboration and social software, while becoming more pragmatic about customer infrastructures. The firm’s “blended” strategy, the word it uses for supporting both open and closed source software, is I would argue more compelling when also seen in terms of allowing customers to mix and match Enterprise Java and Windows, Portals and web-based environments. Remember for example that 50% of JBoss implementations are on Windows (although that balance may have changed since the Red Hat- JBoss acquisition). It would be good to see a deep interoperability partnership between BEA and Microsoft.

Adrenaline: A Shot in the Heart for BEA Portal

When BEA announced WebLogic 9.2, most of the attention   went to the server and Workshop IDE and Integration 9.2. But it could be that WebLogic Portal 9.2 (codename Adrenaline) is in many respects the more important release. Most reporters and analysts didn’t pick up on new functions in the platform, notably RSS-support, so it may be that BEA didn’t give this area as much focus as it could during the launch. I helped foster the silicon.com scoop in that regard, which seems to have generated a fair bit of press coverage (value of an influencer, anyone?)

The relative lack of buzz around Adrenaline could also be because portals are increasingly paper for wrapping fish and chips. What we used to call a portal deployment is now often badged an SOA or composite applications project. Customer references may have signed a purchase order for a portal, but when the media asks they call it a SOA deployment.

Corporate portals can look a bit rudimentary these days, especially in light of the rich experiences we’re increasingly used to in AJAX-enabled web properties. Whether you love or hate the term Web 2.o, it has definitely raised the bar when it comes to rich user experiences, and more importantly, user expectations. Another major change in how we collaborate comes out of the RSS world, through ultra-lightweight publish and subscribe to feeds. . .

So what is the technical story?

With the Adrenaline release WebLogic Portal can now host services held on other platforms, where the Portal can be used as a management framework, through full support of the JSR168 portlet specification and WSRP.

BEA is also working to offer support for web widgets provided by the likes of Google. It is currently in discussions around APIs with the advertising giant, because there are no standards in place.

If blogs have taught us anything though its that sometimes rudimentary technology can be very effective. RSS may be primarily serving text, but text offers a great signal to noise ratio. SMS is another example of text giving us what we need.

Technology is only powerful in as much as people actually use it. Plumtree was adept at was helping its customers build communities around its platform. Plumtree wasn’t about a technical sale, but a social one. You can take an end user to water, but you can’t make them drink. Plumtree helped its customers to build virtual water coolers, where they didn’t just drink, but chatted, nudged through corporate hierarchies, and got things done.  If BEA can harness this social IQ it will be in good shape as the world increasingly moves to socialised business process approaches. BEA needs to harness as much social energy as it can as it diversifies, given that it will need to reach out to very different constituencies from its traditional architect and developer audience.

Tags and Graffiti: Making a Mark

I haven’t seen BEA’s new tagging and metadata infrastructure, codename Graffiti, yet, but the concept is very interesting. What if you could tag any business application by pointing at it, and adding relevant metadata, and then navigate the app based on the user-generated pathways?

IBM is also currently developing an enterprise tagging system called dogear, but that seems more URI-based. Graffiti on the other hand, is being designed to tag non URI-based data structures.

Towards a microService architecture 

Given that BEA is now increasingly a three or four trick pony, partly given the resurgence of Tuxedo in customer shops, it makes sense to reapproach the potential range of services and APIs provided by the firm. Rather than building silos and monoliths BEA wanted to be able to offer customers more choice across its portfolio. It is therefore establishing a new kernel that runs across WebLogic, AquaLogic and Tuxedo, and manages service interactions between them, under a strategy called SOA 360.

BEA has learned from Eclipse here, and is building the microService architecture on top of the Open Grid Services Infrastructure (OGSi). Put simply OSGi is the stuff that enables Eclipse to work well from a modularity perspective. Now commercial software firms are learning from Eclipse, and bringing the approach back into their portfolios. IBM has done so. Now BEA too. Cote says Apache has an OGSi project called Felix, too.

The BEA microService approach should allow for greater customisation in customer shops, and allow BEA to componentise and recomponentise its tooling to support a wider range of use cases. Thus, for example, BEA is increasingly delivering microplatforms that include vertical industry functionality, for claims processing, say.  

The Boring Acquisition Question   

Oracle’s recent feeding frenzy, which took in JD Edwards and Siebel, and some niche middleware (Collaxa) and security (Oblix), surely makes it less somewhat likely it will try and acquire BEA. [since I wrote this Oracle also acquired content management (Stellent) and a utilities software co (SPL Worldgroup).] On the other hand Oracle’s purchases seem to be accelerating not slowing, so its hard to call.

The fact is there are very few companies out there that could afford to acquire BEA – its pretty big in its own right now. It has more products than ever. Customers signing up to new licenses. An entirely new market to tap (and a rather large one at that) called Windows.

BEA is playing the independence card, and playing it well. Sure the core app server market is increasingly commoditised and saturated, but BEA has many more horses in its stable now, and not being part of an application stack is a good thing in this regard. It plays SOA as the bridge.

A Question of Balance: the new see saw

It used to be that enterprises licensed BEA software at least partly in order to keep IBM honest. It certainly wasn’t just a technical decision.

But I am noticing a new pattern emerge in Europe amongst my architect contacts. That is, these days organisations that moved many of their workloads from IBM to BEA, are now in the process of a new transition. Increasingly its a BEA/JBoss seesaw, rather than an IBM/BEA balance point. These days, in many European IT organisations at least, JBoss plays the Amdahl mug role to BEA as “IBM”. This is actually good for BEA- it means its the incumbent, and customers want to exert pressure on license fees. Better that than to be on the outside looking in. Before proceeding I would like to stress that I am not saying IBM is losing share to BEA and JBoss in Europe, but that we’re seeing some new ecosystem dynamics emerge.

Conclusions

I could be completely wrong. Oracle might bid for BEA tomorrow. But customers are still more than willing to pay BEA for its software, and the portfolio is getting more interesting. We’re at the cusp of a revolution to harnessing collective intelligence, where all applications become social applications. That’s an opportunity for BEA. I am certainly looking forward to learning more about Graffiti and customer use cases, in that light.

I think BEA needs a more clearly defined rich client strategy to really benefit from social business, but its making some good moves in that regard led by people like Skip Sauls.

And on that note I really need to get out of here for the weekend.

 

disclaimer: BEA is a client, it paid my T&E for BEA World 2006 in Prague.