Views on the Industry

RO

http://www.aspstreet.com/forum/d.taf/id,22570/fid,270

posted by douglas dooley @ 3:33 PM 1 comments

10/28/02 - 10/28/05

It's been three years since we finally put the ship facing the right direction:
http://www.sun.com/smi/Press/sunflash/2002-10/sunflash.20021028.1.html

It's been three years of hell for me, trying to fight shadows, but I feel like I did the right thing, all along. For those who know me, and know what I've been through in the name of an application server product, you know I will be back. I don't hold any remorse, but on the same point, I don't hold on any support for factions that I once supported:
http://douglasdooley.blogspot.com/2005/07/j1.html

But this should not be confused for support for the quoted Sun person in the P.R. On the contrary, if you have seen my other posts, I am deeply critical of the Java Enterprise System and its master:
http://douglasdooley.blogspot.com/2005/06/oss.html

I will take this opportunity to announce my skpeticism of the Sun leadership. That consists of two people, but filters down to several people who are perpetuating the disconnect between what Sun's customers need and what is being offered. There is one need, and that is a fully functioning platform on AS 8.1, EE with a Portal and JBI and LDAP/WS-I registry functionality. But this is for the shareholders to resolve, of which I am no longer one, as I divested once I left so I had no interest, and could be an analyst. This is a long-term issue, and getting rid of Scott will only be one step in organizing a competitive company. No Blades, no services, and no telecom database (yet). All of these things are outstanding, and all are under-stated to the "Red" item that is AS 8.1, EE. But I have said it all before, Sun has Glassfish, JBI, and EE.

And because of these moves (althought my friend, PY would disagree:
http://www.theserverside.com/news/thread.tss?thread_id=36129
cheers to MB:
http://www.sun.com/aboutsun/media/bios/bios-bauhaus.html)

WebLogic and WebSphere are dead. They do not offer any reason for a customer to spend on licenses. Although Marc is doing a terrific job, I have to disagree with his early pronouncement of putting Sun's app server in with WL and WS:
http://jboss.org/jbossBlog/blog/mfleury/2005/06/06/SUN_open_sources_XYZ_who_cares.txt

There are now three app servers left, after 40 odd J2EE licensees and products over the years (remember ATG http://java.sun.com/j2ee/1.2_compatibility.html):
Glassfish
JBoss
Geronimo
http://douglasdooley.blogspot.com/2005/06/glassfish.html

And the battle will be on for customer adoption with JEE 5 from 2006- . Project Fusion causes one to wonder whether there is a viable software vendor left in the market. Microsoft killed the one that mattered in November of 1998, and AOL/Sun remade it iPlanet, a Sun-Netscape Alliance:
http://www.timewarner.com/corp/newsroom/pr/0,20812,666196,00.html

Ahh, history. I am here to continue the march on Santa Clara, that will go from GRR to NYC to ATL and then out to SFO. I firmly reject the principle foundation of the October 26 announcement that JES 4 is viable, whether or not it is truly available:
http://www.sun.com/smi/Press/sunflash/2005-10/sunflash.20051026.2.html

I support the efforts of the JWS team, even at a time that Marc sees Dell as more relevant to his ambitions. I see Sun as the lone systems player that vanquishes a 25-year old misstep by IBM and Apple to relinquish developer control to a single company. My partner would see it otherwise, and that is why there is WS-I. But the fact remains that Java is being discredited even on its home turf of applications for components that build web services. What can be done?

JBoss alignment on components will provide a platform. .Net integration will come following this. And then there will be a new software vendor to match relevance to ORCL, while chasing and ultimately surpassing MSFT. How will this be? With a unified front against all Java-naysayers, including those that want to keep .Net tightly coupled away from the enterprise market. I would be remiss in not suggesting that there is an initiative under-way to begin this work, and Scott gave away his leadership role to Marc and ultimately to JS (RS, potentially?) to do this. I salute more the trench warfare that is being undertaken by MB's crew of PP and the like, wherever you may be (CC, included: http://conferences.oreillynet.com/cs/os2004/view/e_sess/5820)

But there will come a day when the app server war transforms in to a business war, and the victor will be richer than Google. Let us hope, and let's all plan that that victor will be neither MSFT, SUNW, ORCL, or IBM. It is time for the transformation of an industry that was blunted by .com madness (why project pets.com when you have Model N?), to turn in to a business industry of its own. The Java Economy will one day reach the pages of Time Warner publications and WSJ editorials, but first there will have to come Astro...

posted by douglas dooley @ 8:02 AM 0 comments

Oracle

Who is the real enemy of Java web services? Some might contend Microsoft, in a bit of eponymous irony some would argue that it is BEA's Workshop tool, but I would like to posit that the real inhibitor to interoperability is the company that would like IT to believe that it is solving interoperability problems. Even Microsoft has Hailstorm roots. But Oracle, what have they done for the Java web services market? SOA has bailed BEA out, even though they have nowhere to sell, except with Portals. But in the very least an ESB is in the offing. Their disregard of JBI, notwithstanding. Sun has JBI, Glassfish, and AS 8.1, EE. JBoss has the complete solution. Oracle has nothing. They are a shell of the company that once looked to build functionality with enterprise Java. In fact, have they ever been in the game - - some observers might conclude that they are nothing more than a proprietary vendor looking to compete at Microsoft's game.

I would like to point out though that MSFT has changed the nature of their competition, and even though it is time for Mr. Gates to work on a new project, I believe they are on the right track with .Net and JBoss support. Nothing will be bigger in '05 than Sun's x64 servers, and MSFT's announcement with JBoss, irrespective of Oracle. And so we turn to the database vendor, aspiring to be controller of corporate IT, and back-end operations. What would make you feel that there is a chance that they will play the JWS market. They buy Siebel as soon as Siebel is about to get off the ground with their component offering, they seek to kill salesforce.com when they are the leader in componentized software, and they bury PSFT when they are making a play for JWS support via WebLogic, and potentially even JAX support. Why? Because JWS is a threat to their proprietary infrastructure, and they stand to lose the most in an inter-operable world.

Only BEA can bail them out, and have a two-sided story: one built on Project Fusion for maintenance customers, and one for WL for SOA support with a Plumtree front-end. But until then, Oracle is an impediment to progress, just as MSFT has been for years. It took a power grab threat to force Gates to do .Net, but now he is on board, IBM is on board, JS is on board (get the stack put together), and JBoss is on board, only Oracle stands on the outside, making plays that only belabor the ultimate decision of a BEA-Oracle alliance, though they both live in denial. Until then, the two largest Java vendors for the past ten years are the ones that stand in the way of developer freedom, and inter-operability. I'll say it to be crystal clear, only portability brings value, and this is accomplished with AVK and JAX, until vendors sign-on to this, the software market is adrift. Glassfish puts the pressure on. JEMS puts the pressure on. Biztalk puts the pressure on. But only Oracle can make the fateful decision to play by the rules of the game. Until they do so, and if RS is calling the shots, it will be some time, the market has a lot of problems. I would guess it could take 10 years to sort out.

In the meantime, Oracle will continue to buy threats. And the great hope will come from start-ups that take shots at the "Java" giant, while never quite having the solution. It will take a great start-up that can overcome inertia and the slide toward dependability on database lock-in. Chuck Phillips had a plan, and he is executing it, but since when does a financial guy have the developers' interests in mind. And does he really have control of the process. Who is actually running Project Fusion, what is its intentions. Is it a threat to J2EE? Why is BEA against JBI:
http://jcp.org/en/jsr/results?id=3226

I ask these questions in the hopes of finding out what really lies behind the intentions of the biggest Java vendors in the marketplace. I want to know whether they support the Network as a Computer, or do they support only the interfaces that maintain lock-in. We struggle with these questions in every industry, in every company around the world. The answer is universally true - - standards matter. And as long as Oracle and their little sibling, BEA, continue to play a different game, the more prolonged the JWS market from happening occurs. And the longer the business world must wait for the World to be Flat...

posted by douglas dooley @ 7:56 AM 3 comments

step 1

http://www.sun.com/emrkt/x64gridrack/learnmore.html

posted by douglas dooley @ 6:36 AM 1 comments

Google

The big news out of CA is actually fairly substantial, though all of the would-be technology journalists thought that you could fit 600 MB down a 100 MB pipe. Not yet. I found the announcement of Sun-Google significant for one reason: kill JRockit before it gets any footing. Not that that would provide BEA with any room to maneuver on the OS side of things, perception is king, and as Mr. Doerr will point out to any classroom, "All Technology is Political." The Google arrangement makes Storage Tek look like it may have some room to maneuver, for it substantiates the philosophy at Sun to be an Internet arms dealer, and provide companies that have Extranet capabilities, such as Google, with the ability to integrate the whole Computer (with a big C) together. Google is not a web services vendor any more than Hailstorm is a web services offering, but one can only imagine what good things ould happen next.

I, along with the rest of the digerati, speculated on what yesterday's announcement could be. A Becholsheim push would have been irrelevant unless Google was prepared to standardize on the boxes, which they will not do. They could have announced a "thin" Star Office, though even JS disbanded this speculation, among his pontifications of Java and services and the web community and all other responsibilities that he had yesterday, managed to set the record straight that Star Portal was and is D.O.A. No, what I thought would be wonderful would be some investment in SJS AS, EE - - - of course. Now that would have had some pull. By saying that the largest transactional web-site was standardizing on Sun's Java OS, instead of just re-distributing Java code, you would have had the makings of a Java economy built on Java web services (this is my game, JS and RS), and then componentized Google services, and .....uhhhh. It was not to be, not even Eric Schmidt could pull off such a coup that would actually threaten the MSFT establishment. Re-distribution of browser code is not the same as distribution of JWS code, and web services is still something that Gartner talks about to fill Orlando hotel conference rooms, while business goes about renewing Oracle licenses (that is Siebel and PSFT maintenance), and not seeing the real opportunity posited by JAX, Glassfish, and AS 8.1, EE. I will throw in JBoss as a hedge.

No company can compete with Microsoft today because no company is willing to go in to the den where it means the most: pre-built functionality, to combat VB. Not even Google is willing to standardize on Java. They'll re-distribute it, but they won't run it. What's the deal? The deal is a JVM that is withering on the vine while we wait for Niagra and talk about 4U servers, which is all good, like Stroage Tek is good, and Open Solaris is o.k., but it does not a single thing to execute on the opportunity in front of Sun, and ultimately in front of Google, which is an inter-operable marketplace built on auction functionality that allows developers to bid for their services with firms, and allow conduits to make money off of the transactions. What is Google's market cap.? Because of Advertising ! ? ! They are not an e-mail company, or a Blog company, or a marketing company, they are an enterprise software company with reach in to the holiest of grails, the users within corporations, and a dead lock on its developers. Monetizing developer adoption in the form of a Java web services marketplace is the only way to climb the mountain that Netscape came close to doing, and which Sun can arm them for. Google needs to develop marketplaces that do not compete with Wal-Mart, but rather competes with ComponentSource. Then I get interested. Until then, it is just a marketing revenue game, with little to no strategy for being "relevant."

I admire Scott's tenacity, but maybe just maybe, he is in the way for Google to go deeper with Sun. We all recognize that he will be gone within three years, if not three months. I just hope he can go out a leader, as Eric said, and take the bow, and look for NC's to come back as Act II. Right now the battle is over infrastructure software in a realm that he is not familiar, and which his number 2 has a lock-on (reference yesterday's smoke blowing by JS). I think there is a lot to be said for what happened yesterday because of the enterprise software implications, but if Sun, Scott, Eric, or Google are focused on hardware, JVM's, Portals, or monetizing audiences, then there needs to be a change. The future is in JWS. End of story.

posted by douglas dooley @ 5:00 PM 1 comments