Views on the Industry

pre-built EJB components

The case is simple, and it has a history: build functional applications that are cross-platform, and revive the model of Java to be write-once-run across multiple platforms.  This was done, in some form, by Theory Center before they were acquired by BEA in 1999, and it allowed WebLogic to take the early lead, that ultimately led to their sizable cash-out, in the form of the Oracle acquisition.  Nothing made IT managers happier than to see something actually working on top of these Internet operating systems, that were to be known as application servers.  They could see what they were investing in, if Theory Center worked, then their own development efforts could also work.  I was working on iPlanet's competitor to WL, and never once did I see a deployment of Theory Center, yet many of the largest financial services firms and teleco's, deployed WebLogic in their environments.  Some will say that application servers never lived up to their promise of integrating all of Java, in to one common platform, that could rival Microsoft's datacenter offerings.  But that is what is available today: a belief that cloud computing will re-orient the data-center like no other offering brought forth by enterprise vendors in the last twenty years.  It is the Internet taken to its natural destination, and it is custom-made for application server vendors to take their offerings in to the cloud.  The vendors, themselves, are coming at from different angles.  JBoss is the lead environment for Red Hat Linux, and OS that desires to see the market go cloud, it is Oracle that will lead with applications for intra-environment deployment, and IBM will offer up some customization to argue for their own tools to be deployed as cloud infrastructure.  Even Spring Source is in the cloud game with VMWare having invented the enterprise cloud architecture.  What do all of these have in common?

They all are lacking the Theory Center moment, when you can walk in to a customer and visibly demonstrate what is being explained in technical terms.  The promise of pre-built functionality has always been on the minds of the vendors, and customers have their own in-house set of components that run certain functionality, and that can be re-used across multiple projects.  What is needed is a firm, a coalition of firms, or a series of start-up moves to re-kindle the fire originally set by the promise of the EJB component model.  Pre-built EJBs are what would give the cloud that discernible advantage over the Azure promise of integration across the platform.  Windows Server is acquiring the enterprise, slowly but surely, with 50% of deployments worldwide having gone to WS, and 20% with Linux, and 30% other.  In order for the remaining non-Windows 50% to compete, it needs these pre-built components to show the power of Java's development model, and work across the vendors to prove the deployment model.  Whether that be IBM Global Services, or Oracle's Support Network, or JBoss consultants, worldwide, the component model needs a new life.  This can be done in basic steps toward a full-fledged platform, that would deliver shopping cart, credit card transactions, other shopping features for consumers, B2B transactions like supply chain management, integration with all sorts of data sources are ready for pre-built functionality, and beyond.

The best way to deliver this is on top of Glassfish, as the Reference Implementation for JEE 6 and 7, it can showcase the advanced cloud features of Enterprise Java.  Pre-built components would give customers something to work with as they investigate the business model of going with Java over Azure in the cloud build-out.  Microsoft will be forced to retaliate, but they do not have an industry that has committed to delivering specification after another, with partners, that compete for the same accounts.  The Microsoft components will be more centered at Google, anyway, their Great Plains product-set has been focused on ERP and other large functionality efforts, not micro-enough to be componentized in time for Java components to build a user base that will set them off on the cloud projects, that are certain to spring up, as IT managers look to new models to improve delivery of their respective employers' web offerings.  Get EJBs back in the discussion, and integrate with other non-EJB models through JAX, and see the application server vendors regain the argument that has been eroded by years of over-hype, now is the time to deliver on the hype of Java on the server-side.  Right now, it is all a process of upgrades and maintenance decisions, make it about the cloud and see new models and new opportunities rise to the surface.

This cause has been called for before, it has been dis-credited, and to be honest, is like kryptonite to IT everywhere.  The promise of pre-buuilt components, a long-time Holy Grail in development, is a real possibility in deployment, today, with the clouds becoming the leading selling point of vendors, and the leading interest area of customers.  How to make these clouds work?  Microsoft has it figured out with their complete package of development and integration among a wide-swath of products.  All Java needs is a development environment, pre-built components, and application servers to compliment the data sources and integration efforts already in place across the enterprise.  There is no need to sell new platforms, or new models, the app servers are the cloud OS, always have been, they were just disguised as dot-com enablers, in their earlies incarnation, but they power the Internet.  Where would IT be without enterprise Java, it would be stuck in non-standard, non-compatible environments, without any hope of achieving acquisition or other-wise integration, it would be much costlier to run an IT environment, without application servers.  Re-charge the debate around Java with pre-built EJBs, that work across cloud environments, make it a competitive marketplace for selecting components, that can work together, and turn the power back to developers within enterprises.  The cloud will shake out from there.  Without it, there is no functionality to sell on, and allows Microsoft and Azure all the advantages of integration without standards.  All of the Java vendors would benefit, and Glassfish would have a standing chance of becoming one of those deployment environments ready for the cloud.  This is Oracle's best opportunity for the cloud, beyond whatever WebLogic does to become Fusion for the Oracle Cloud, a pre-built set of Java specific functionality would be a major starting point for how Java vendors are to approach the cloud.  It would translate to benefits across the Java ecosystem.

posted by douglas dooley @ 12:58 PM 0 comments

Sun software spin-off

If its not going to be used to sell hardware, then it should be given its own life, outside of Oracle, its wasted value, just to be specification Reference Implementations, that is comical, its industry-grade, and can stand on its own, it should be set free.  MySQL, SSO, ESB, JEE, even NetBeans could sustain a whole new company, and these are redundant assets within the parent company's offering, they are not investment worthy, so why not let someone else do something with it.  Cannibalization is possible, but considering the pay-back on green environments it could get to and turn-back in to maintenance revenue with a simple agreement from parent and spin-off, there is very little logic that just letting them wither on the vine in the argument that Sun was bought for hardware alone, and it is in keeping with the long-term principles of holding off Great Plains in medium-sized accounts.  I have said this all before with some degree of implementation plan, like my earlier logic on a fork on Google Code, i dont need to see that exactly happen, but it needs room to breathe, and right now the oxygen on Glassfish, in particular, is being cut-off from ExaLogic sales push.  This is not the only value of Sun, more can be done, and the only way I can see anything getting near its true value is for a spin-off.

This reminds me a lot of iPlanet.  Back then, when the Sun-Netscape Alliance was in its first year, there was a lot of hype for the Sun investment in Netscape products.  I was a kid, on my first job out of grad school, and actually advocated for an IPO, i was naive, but eventually fell in to line that it would work better as a vertical solution.  And it did, it served Sun well, even though it caused us to work over-time to clean up the iAS 6.0 mess, it served Sun well.  That didn't show up in the acquisition price, but Sun would have been virtually a worthless chip designer had it not been for the middleware assets.  Oracle is doing its best to revive Sparc, and i applaud them for that, I think they should continue doing it, its great for the Valley, great to overcome Google's black-hole of data center design support, other companies need non-Google cloud data centers, and Oracle with Sun hardware fits the bill.  Good job whoever worked on that, even Solaris seems to be safe for existing accounts.  But the other side of Sun is breath-takingly innovative.  All the middleware assets are better than JBoss and Spring Source, and by not addressing their value, Oracle is giving legit competitors a little more breathing room.  Just keep 51%, split up 49% with investors who are savvy on enterprise software, and let some employees have a period of time where they could be employees of Oracle, still, while working on the start-up of acquired Sun software assets.  Its really straight-forward, and is custom-made for an aggressive company like Oracle.  It would appease the Sun employees who are not fortunate enough to work on Fusion, and give a little energy to the growing international monstrosity that is Oracle.  It would simplify things for everyone.

Customers would know that they have the support of Oracle, if they invest in this new iPlanet-like company.  Call it the Oracle-Sun Alliance, or whatever, just get some value out of it.  Its like Oracle executives would rather keep WebLogic as the only option to punish customers, they are eventually going to get something better in IIS, .Net, and Great PLains, when it is all rolled-in to one environment on VS, only Glassfish stands in the way of this happening.  Dont let it die out of fear of the unknown, like WebLogic will be dead if there is some innovation, WebLogic is not going anywhere, it is the development environment for BPM on the ERP apps, that is huge, like really huge, but it cant be expected to also fend off more nimble cloud purveyors as well, that is not logical.  Oracle needs both, but that is not what is happening now, customers see the messaging about departmental apps, so no matter what Glassfish people tell Oracle sales reps, there is no way to get in the customer conversations, it is as an after-thought in Oracle strategic direction.  There is no reason this should be the status quo, there is room to move forward, but it takes fresh thinking, and see the challengers as real, dont get caught flat-footed on your flank-side.

posted by douglas dooley @ 1:02 PM 0 comments