Java Economics
How important is portability? Since the release of J2EE, vendors have claimed adherence to standardization, while telling customers how to implement special features only available on their platform. But beyond specification compatibility, what is portability and how does it relate to developers?
First, portability is simply the ability for applications to be deployed on two or more app servers. Second, the measurement of this rests with Sun's Application Verification Kit (AVK). This is all that needs to be known about portability. If a vendor does not provide a standard run-time, then it does not support the value proposition of portability. If an application does not deploy on two or more app servers, then the developer has chosen not to adhere to J2EE standards. The implications of non-portability are nearly always higher maintenance costs, whether or not the customer re-deploys on a different app server at any point in the application's lifecycle. To avoid non-portability, customers should use the AVK to test all of their J2EE applications and components, and take steps to correct for non-portability. If the AVK certifies that the application is 100% J2EE, and it can be determined that the app server is non-compliant, then the customer should switch platforms. If the application or component proves to be less than 100% J2EE, steps should be taken to make the application compatible, and thus portable.
Developers, more than any other IT constituency, understand the benefits of portability, especially as it relates to maintainability. It cannot be epected that the same developer will manage the app throughout the lifecycle, and therefore developers monitor portability to enable other developers to make changes and enhancements. However, the best means to ensure portability and save customers many hours of development work is to use Sun's AVK.
What are web services? Java web services defines the specifications necessary to enable inter-operability of applications and components. The specifications to allow Java applications and components to be exposed as web services are the Java API's for XML, known as the JAX technologies. The JAX technologies to map, bind, register, and issue remote procedure calls on web services. JAX became a standard implementation with J2EE 1.3, and will provide additional capabilities with the release of J2EE 1.4. The key point for customers, developers, and system integrators is to get started with utilizing the ease and flexibility of Java web services by learning the capabilities of JAX.
Terror Incorporated
Loretta Napoleoni has recently issued the definitive analysis of geo-politics thus far - - Terror Incorporated:
http://search.barnesandnoble.com/booksearch/isbnInquiry.asp?userid=tI5O4IpBb3&isbn=1583226737&itm=1This book has made the case for a worldwide economy of underground organizations, which should come as no surprise to those tracking current events. But the extent and the sources of this system are revealing, and should impact U.S. foreign policy, especially when Americans see what has been hiding below the surface for some time. This is an industry I hope to learn more about and write about in this area. At this time, I will stick to what I know and have written about the Java economy, and leave it to the experts, such as Ms. Napoleoni...
Manifesto
It's the dead middle of summer, but this was originally written on December 31, 2003:
"The impact of interoperability has not been fully explored. The Economist provided some overview of the reach of web services integration in their April 2001 survey on software, but failed to truly describe the constituency where the forces of change are multiplying exponentially: the development community. Whether or not Microsoft follows the push for true interoperability is all but irrelevant. Even the most powerful company in the world cannot derail what began as a marketing program. The J2EE market is developing in to a full borne Java Economy with enough resources to alter the way business is conducted. The shift is described by Shoshana Zuboff's thesis: The Support Economy
http://search.barnesandnoble.com/booksearch/isbnInquiry.asp?userid=tI5O4IpBb3&isbn=0142003883&itm=1
The rise of distributed platforms of commerce with services to facilitate inidividual relationships plays in to the hands of the J2EE developer community, where distributed is the name of the game and services are the trade. The estimates of the size and reach of this community do no good in understanding how far this paradigm will extend. The nature of Java has been one of adaptation to the opportunities presented by the marketplace and the community of which it serves. There is almost no exception among the IT vendors that offer development and implementation of Java services, all except the most powerful one.
At the other end of the spectrum is a start-up, the first true Java start-up that has demonstrated a sustainable business model: JBoss. By unifying developer productivity with developer control, JBoss has re-oriented the Java OS market in two years. After surviving like "cockroaches" through the investment "nuclear winter," JBoss out-maneuvered the application server vendors that were noisily building 'solutions.' Now that the core Java run-time is controlled by the open source community, the rest of the Java web services market must play by their rules. These rules have resulted in the elimination of licenses revenue in the run-time, and has focused spending on the services which run on the app server. Even the 'core' functionality of portal, integration, and development environment fails to differentiate the other vendors beyond their core run-time.
Now customers have real choice and can make infrastructure decisions based on whether their services needs are available or whether they have to build them. All other functionality is secondary to what exactly is going to be operating within the environment. The customer has also regained the leverage that was lost in the great WebLogic hype of '00-'01, where business deals eliminated the vast majority of IT budgets before any business application had even been scoped. Now the reverse is true, where customers can choose from IBM's WebSphere Global Services, JBoss Professional Open Source, Sun's Java Center (and similar offering from Oracle & BEA) to fulfill their architecture requirements. All the spending has shifted from infrastructure to services with best practices and reusable code providing competitive advantage. With this model, the value of the engagement can be measured by the resulting output code for services. Now that the vendors can no longer get margins on the Java run-time, the competition over all other software and hardware spending is commoditized away.
The only competitive market left is owned by the developer community, whether the vendors are in a position to admit this or not. The last remaining vestige of IT competition is software services and the leadership position belong to JBoss, IBM, and Microsoft as the most adept cultivators of the developer community and/or pre-built services. But more importantly is who is going to emerge to fulfill the capabilities of the distributed platforms that customers are moving toward with every new Java implementation. All other investments of IT budgets decrease the agility and thus increase the time required for an organization to fulfill their new operating environments. The companies that move their Java run-time to JBoss and begin building the services for their distributed relationships will organize the initial "Federations" where the next generation of economic development will arise. Only by orienting toward a Java Platform with Java web services interoperable across the extended value chain will the promise of distributed capitalism take place."
(This is dedicated to the Sun application server product managers who continue to get no respect, though keep the Java Economy afloat - - djd, 1/5/4)
Coming Next:
a. Java Economics
b. Java Implementations
c. Distributed Platforms
d. Developer Markets
e. Operating Federations
f. New Business
MSFT
Good afternoon, hello, it is Friday again, and time for a new post. Not much has happened this week, pretty much the lullest week of the year. I would hope that the real big news of Microsoft being at J1 will actually amount to something. My bet is that there is nothing of substance to come out of the great alliance for another year. What more can be done? I think they have pretty much taken care of the directory issues, and the only thing that would be of consequence would be an improvement in the JAX-.Net story. So J2EE and the .Net Platform would become interoperable. That would mean something. But that would require certifying .Net SOAP with the J2EE RI, and then you would have an unfair advantage for GlassFish. How could the agreement to settle with MSFT not include this? Well, it is only because when you negotiate with MSFT, you are most likely in a disadvantaged position, as management was, and therefore you think that doing something with a dead product like Directories is worth the future of your company. I think it is not. Now you have Microsoft running around the world claiming interoperability, because Sun says so, and there is absolutely no proof that they are. Name one time in the history of the empire that it has actually kept its word. It has never happened. But McNealy and team have bet their survival on interoperability. It's a joke, and a costly one. Gates is an impediment to moving the software industry forward because he has a maniacal mind bent on domination, nothing has changed except his wife has decided to give away some money. I thought Ellison's move to fund an analysis of when money is spent correctly on public health is great, and should put a stop to Gates thinking he has carte blanche on world issues. Microsoft is a great company, but it is run by a man that smells blood everywhere because he has created a bloodbath, and I am personally offended that he has taken the fight to Sun. But the most offensive part is that (literally, unbelievable) Sun has gone along with it. It would be one thing to agree with Dell on blade specs. so that there could be some interoperability, that would actually be a good move for Sun. McNealy sees a fight everywhere, and Gates sees a victim everywhere - - they both should go. Their era is over, and it is time to pass the reigns to a new generation of leadership that can put the in-fighting behind it.
Next on target: JBoss. The one company that threatens everything that Sun-Microsoft is trying to accomplish, they represent the convergence of OSS w/ JEE, and there is no way for the two of them to control either of them. GlassFish would be interesting if it could ship sample apps. and ISVs on the .Net Platform, but it will not be able to catch JBoss in the medium-term. That would only come from a commitment to Solaris w/ AS, EE, but for whatever reason that does not seem to be a priority. As we have talked about, Sun has all of these initiatives but no coordination. But just because Sun can't get to them, does not mean that Gates won't try to get to them. It must be some kind of rush to be in the cross-hairs of that man, especially when you have him beat. Sun is handing an entry in to the competition by thinking that Liberty is for some reason relevant. Microsoft will gladly accept a life-line to their fledgling middleware strategy. No one is doing .Net, and why would you. Ultimately, there will need to be an argument for doing .Net infrastructure over JBoss infrastructure, but until that day comes, Sun will seek to bolster the claim that .Net is relevant. It is not. It has not shipped anything of relevance, everybody that is doing VB continues to do VB, and VB.Net is just a piece of some puzzle to make it seem relevant to its constituency. Until there is some demonstration of SOAP integration via JAX/.Net, then JBoss will not have to be a Java web services company. And as long as that goes on, BEA gets to claim leadership in a nonexistent market via Liquid. Sun has screwed all angles up by not tying in J2EE with .Net, and yet MSFT gets to claim that they have tied in with .Net, because that is the kind of lingo they are familiar with (Sun signed an agreement with us, so now we are compatible). I am tired of pseudo-legit business deals going on when the developers keep Sun alive, and Microsoft hopeful. Its almost time for a Manifesto, I think there needs to be some statement of what is needed, and what the plan is. I don't know if I am the appropriate person to do this, but I think the time has come to provide a base for what the future holds, and this time, it is one without the guidance and direction of Sun Microsystems...
J1
Ahh, as I kick back after another week, I wanted to reflect on the great event of JavaOne. Already wrote about the big GlassFish news, and JBI continuing seems to be a once a year announcement (the great luxury of being an outside observer, is to do what so many young Java developers do when it comes to project initiatives, and that is to generalize). So what big news is there to discuss, well it's an area that I never thought would be broached by my writings, but I need to remove any historical bias and state what is the next most logical move for Sun. It has been a long time coming, as the last four years have seemed like a decade (n'est pas?), but for many reasons that I will now outline, I have come to a conclusion about future direction. One reason is the most obvious: MSFT. After a decade of vitriolity, Microsoft is now a showcase representative at the great show. What does this mean? Well, to Netscape people, this means the next logical extension of web services influence in to the hearts and talents of the Java community. How soon we forget how close the Netscape middleware stack came from supplanting MSFT in the Internet infrastructure war. Now because I joined the S-N Alliance the same month that NAS 4.0 was released, I know that it had a major impact on the launch of J2EE, as iPlanet Application Server 6.0 was the first J2EE shipping product:
http://www.sun.com/smi/Press/sunflash/2000-06/sunflash.20000606.4.htmlMarc and BEA will say how it was either irrelevant or not surprising that Sun pulled it off as the first, but remember guys who pays the bills, its that little logo that I got through legal in May of 2000, so ante up if you want to talk about size. Back to the discussion today, I believe in an interoperable world with .net, but I am not sure it should come at any cost, including membership in to the Java community. Secondly, the world of enterprise computing has moved dramatically away from the hardware specs. that have defined Sun and the analysts are right that what is going to define Sun is a software infrastructure that is only now being assembled. SeeBeyond is a step, JBI is a step, Glass Fish is a step, and the Sun DB (read: Clustra) is the next logical step to bring this to a competitive level. And I am not wholly critical of StorageTek acquisition, I think it was an interesting move for financial stability, and to theoretically integrate in to a cohesive solution for selling in to storage accounts, whether or not it will ever integrate with Pirus or other virtualiuzation efforts is a different blog. (the best thing about blogging as opposed to writing in Word, is you don't have to be told that your sentence is too long without a suggestion as to how to make it shorter - - damn that is so condescending; I know, I know, i need StarOffice though that probably does it as well, and the corporate world just isn't there yet) Initiatives abound at Sun, very little is actually being accomplished.
I guess what I feel is that the best positioned company in the industry continues to make incremental progress. After years of over-hype, there is now under-hype, and that impacts the bottom-line of a lot of outlets that count on the resources that Sun brings to the market for selling solutions. The current focus on data center sales ignores the realities that all other IT firms have embraced, and allows a company with no expertise, like Dell, in to a possible winning market at the expense of the developers. Nothing was done at JavaOne that made the Java community any better off, maybe there was PR for Sun, but there was not an initiative that brings direct benefit to the developers. I want to see a force that will open up business for the developers, like commercializing java.net or creating a venture fund for Java web services applications, anything would be better than the ONE campaign for human digitalness. I guess what I am finally coming to is that the company is adrift, in some similar way to how the U.S. policy in the Middle East is adrift. It seems good on paper, but is just not working out in practice. Why is this?
Well, there are a lot of ways to look at it. The side that, for example, GWB comes down on is the glass half-full side, where Sun is starting to make money and has stopped the bleeding, but is no where near a growth strategy. The side that I look at is that something is amiss, there is something very wrong with the company's direction. Not least because it has some of the greatest assets, no let me re-phrase that, the greatest assets in the business world, and nothing to show for it, other than the constantly upticking of new developer adoption figures: 4.5 million strong as of this week. I want a company that will go beyond what is feasible and go after what is bigger than that. Too many people are invested in Sun to see it become another mid-tier company in the Gartner quadrant analysis - - it should be dominating every category whether or not Gartner is irrelevant. Gartner should be made irrelevant because Sun dominates the IT categories (I'm so sick of Liberty, N1, Open Solaris, and AS, EE being merely skunk work projects), and so there is no need for the quadrants. Instead, it is behind in every single category, and claiming that the problem is Global Services. The problem is under-delivery. Almost every single project on the Sun agenda is Red. There is an execution problem that is fundamentally affecting the entire IT industry. WebSphere is behind schedule because it does not have a direct competitor (no, BEA the world does not come down to app. servers, though when it does, you lose to JBoss anyway); Oracle can name middleware Fusion, and not be called out on it; Microsoft can bury Hailstorm because there is no Java competitor to make them compete. The bottom-line is that it is time for a change. The ultimate decision is for the Sun board to decide on what the next step should be. As a long-time admirer of the man, as a believer in the vision, and as an adament supporter of his status in the Silicon Valley lore, I think the time has come....for, Mr. McNealy to step down...
