Last August I noted that Microsoft regularly finds itself buried under an avalanche of news coming from its Java-based competition. It’s impossible to compete with that kind of volume and that fact alone has caused the SOA market to gravitate toward Java and away from .NET.
Well, something similar is happening this year with open source vs. proprietary vendor in 2008, but, in what should be considered a bit of a stunner, it’s the open source folks who are creating the news deluge. It started innocently enough when Mulesource and WSO2 both released REST-based SOA registries. Then Red Hat released a modularized SOA platform in February. Now WSO2 and Mulesource are back with another major round of announcements. Based on its December Spring Integration release, you can expect SpringSource to become an increasingly visible player in the SOA market. Sun Microsystems will surely have some service-oriented dogs and ponies to show off at next month’s JavaOne conference and Eclipse, which has already debuted the Swordfish SOA runtime this year, will have a whole slate of SOA-enabled tools in its June Ganymede release.
The open source players are pounding away at the news cycle, throwing a steady stream of innovation into the mix. Obviously traditional app dev titans still dominate the market in terms of dollars and customers, but it’s about time somebody noted that we’ve got a movement on our hands. If you’re looking to build loosely coupled services, there are a host of open source vendors to choose from and that ecosystem is growing at an aggressive rate.
Change, particularly in an established market, doesn’t come in one big seismic event. It takes years of consistent pressure to remake this kind of landscape, but to be sure, we are in a period of volcanic activity for the open source market.
Defining your terms makes a world of difference when a project manager is modeling a business process, says Debra Berard, program manager for business excellence, Lean/Six Sigma at Seagate Technology LLC.
The bugaboo that also haunts data integration projects — you say “bill,” I say “invoice” — is something project managers need to solve in business process modeling for application development.
A recent example Berard offered was the design of Seagate’s failure analysis common tracking system (FACTS) application, which is used to find the root cause of failures in product design or manufacturing so they can quickly be corrected.
In a competitive business like disk drive manufacture the quicker a failure can be remedied, the quicker a new product gets to market.
To develop the FACTS application required WebEx meetings and conference calls with stakeholders from all the Seagate facilities involved including manufacturing sites in Thailand, China, Malaysia, and Singapore, as well as design centers in Oklahoma City, Minneapolis, and Singapore.
During these meetings, the project manager captured the processes that existed in the various locations using a business process modeling and analysis tool, the newly released Metastorm ProVision 6.1 enterprise modeling product.
The first thing the analyis revealed was the while Seagate’s goal was to have one failure analysis process, there were approximately 25 to 30 different processes in the company.
But after further review, that wasn’t as bad as it first looked.
“Come to find out, we did have a lot of processes,” Berard said. “but what was revealed was that they were really doing the same process, but calling the activities different names.”
So the issue was resolved in the conference calls by getting all the stakeholders around the world to agree to call the failure analysis activites by the same set of names, she said.
Once that was done a common model for FACTS was created, which then became the requirements document for the $5 million application development project.
Now, everybody involved in failure analysis at Seagate uses the same terminology as well as the same Web-based FACTS application.
Two interconnected debates are raging in the blogsphere over how service-oriented architecture (SOA) may be running into resistence in 2008.
One debate covered this week in a SearchSOA article is about whether SOA is reaching it’s goal of linking IT and business people, or is just becoming an IT-only initiative that is stalling out in most companies.
The other related debate concerns whether the R-word a.k.a “the current economic downturn” will hurt SOA adoption.
These questions were raised when Gordon Van Huizen, vice president of SOA at Progress Software Corp., called this week discussing how his company is putting all its SOA products under the umbrella of the Progress SOA Portfolio.
The portfolio offers a broader SOA marketing message by covering all the Progress products for:
· Enterprise service bus (ESB)
· Business process management (BPM)
· Complex Event Processing (CEP)
· SOA management
· Data interoperability
· Mainframe integration
From a sales and marketing perspective this seems like a good approach. But Van Huizen was asked about the larger marketing challenge posed by current economic conditions.
Does Progress have a strategy for coping with hard times?
“We have this built-in strategy at Progress that is relatively unique,” Van Huizen said.
He noted that his company’s Open Edge development platform is sold through an ISV channel that has been in place for a long time. The 1,500 ISVs target “very specific, narrowly defined segments of vertical industries.”
By throwing the net so wide, the ISVs reach customers in a variety of industries, not all of which are hurting in this economy, he explained.
“So if there’s a slow down in financial services we don’t feel it so much in that product line,” Van Huizen said. This is also true of the SOA products, he added.
“Our orientation to the market was initally around financial services and telco, perhaps because of the messaging orientation of Sonic MQ,” he said. “But with the SOA management product, with Actional, there’s been opportunities to branch well beyond that.”
The Progress SOA Management product is penetrating into healthcare and higher education, two areas generally considered recession proof, he noted.
“As one part of the market goes down others remain somewhat boyant,” Van Huizen said. “So if there’s an offsetting strategy I believe we have one. Of course, if everything tanks, we’re all in trouble for awhile, and that’s just the way it goes.”
As for the other related debate about how to keep SOA from stalling out even in companies that have the budget to do it, Van Huizen suggested two strategies.
The first one is to reach across from IT to business people by explaining SOA not in terms of technology with acronyms, but through business case studies that show the dollars and sense success of the approach.
The second is to begin SOA projects at a tactical level, even as simple as application integration, sometimes called EAI 2.0. This allows IT to show the business managers and executives the advantages of SOA without asking them to shell out big bucks for a massive implementation.
These two approaches are also echoed in a recent blog by analyst Joe McKendrick, who quotes fellow analyst Tony Baer’s view that this will be a year of lower expectations for SOA.
“Recessions tend to discourage the kind of long-term thinking that grand enterprise architectural exercises are supposed to support,” Baer said. “In that sense, SOA has been caught up in the middle – roughly six years after the current incarnation of the concept emerged with Web services, there remains considerable debate as to whether it makes sense to take a project or architectural approach.”
However, McKendrick makes the interesting point: “It’s worth noting that the case for SOA, in tandem with Web services, was forged during the worst IT spending slump in a generation – the 2000-2002 time period. Companies and IT professionals were attracted to the SOA/Web services concepts because they offered the attractive advantage of building or exposing existing applications at minimal cost and disruption.”
This podcast with Mulesource CEO Dave Rosenberg covers the role of the enterprise service bus (ESB) inside an SOA. Rosenberg notes that an ESB shouldn’t be thought of as a singular piece of software sitting in the middle of every application, tossing aside the hub-and-spoke model from the EAI world that often gets grafted onto SOA. He stresses that SOA “is not about integration,” but rather a sensible infrastructure that can handle modular development and changing business needs.
Other topics covered in this interview include:
- How users are more likely to have an “enterprise service network” with multiple ESBs rather than a single or master ESB
- The role of open source in SOA development
- Why neither the Java EE nor .NET meets are well-suited to service orientation
- The roadmap for the Mulesource, particularly in the management area
- What constitutes “SOA infrastructure”
- Data issues, including getting data out of 3rd party SaaS applications
Oracle has standardized on a JavaServer Faces (JSF)-based RenderKit, which allows the developer who has learned JSF to assemble disparate components into a Web 2.0-style mashup.
Enterprise customers are looking for ways to avoid getting caught up in such complexities, so the philosophy behind the tools Oracle has in beta is to automate the rendering technologies, so developers only need to work with components and pages, he said. This approach also is designed to insulate developers from the on-going changes in underlying technologies for RIA, he said.
“As technologies change, we can change our framework but they don’t have to change their pages,” Farrell said.
He describes the Oracle RIA tools as “very WYSIWYG.” The developer designates that a page will be Ajax with Flash from Adobe Systems Inc., Farrell said, and that is all the coder needs to know about those technologies.
Farrell said the Oracle RIA tools are in an advanced beta stage prior to the official release. Interested developers can find out more information and even download them from the Oracle Technology Network.
Recently we polled SearchSOA.com site members on their RIA and composite application plans. What we discovered is there’s a massive overlap between the SOA and RIA audiences.
In all we received 395 responses and 44% said rich Internet applications were part of their enterprise IT/business strategy. Another 30% reported that RIA would become part of that strategy in 2008. 85% reported that RIA was an important to extremely important piece of their SOA strategy. Only 2% said RIA wasn’t important at all to their SOA plans.
Most strikingly, 74% reported they expect the importance of RIA to their IT/business goals to increase this year. In other words, for 3/4 of our survey respondents, RIA is a big deal that will be getting bigger. Rich Web front ends were the most popular type of app being built or planned (79%), with Ajax (81%) being the most popular technology employed to build those apps. Yet 55% also reported they are building/planning database composite applications and 35% reported they have entered or will enter the fairly new space of enterprise mashups. That’s a fairly massive amount for a category that would have been in the low single digits two years ago.
Oddly, mobile apps only drew a 29% response rate. That could be read a few different ways. Our respondents were mostly senior folks in the app dev or IT department. It’s possible rich mobile development is being done outside their auspices. Yet the fact that the more senior people in the app dev arena aren’t connected to it would also mean that rich mobile development hasn’t become a major enterprise initiative. The other way to read it is that mobile devices have yet to become a major business initiative. In fact, mashups using unified communications might be the path that mobile devices take rather than strict mobile app development.
The top two benefits sought by those building out rich/composite apps were improving the user experience for customer facing apps/services (65%) and providing expected levels of business functionality to end users (61%). Lack of internal knowledge/resources ranked as the number one obstacle to adopting Web 2.0 technologies (21%). It also ranked high as a secondary issue (35%). Yet a whole host of issues fell in the 27-38% range for secondary issues: techinical readiness/back-end support, selecting the right technologies, security, data/application integration issues and application performance issues.
Finally, IT management ranked as the top evangelist (28%), technical decision maker (34%) and financial decision maker (40%) when it comes to Web 2.0 technologies. Yet an interesting person ranked second in evangelism (27%) and technical decision making (26%) – the architect. Maybe this has something to do with polling the membership of an SOA site, but it speaks to how architecture is becoming a primary concern in all applications work these days.
It should be remembered that for years analysts have been saying that a primary benefit of pursuing SOA is to get ready for whatever comes next, to be able to deploy new technologies on top of the existing IT infrastructure in a way that makes sense. It would seem from our survey that those predictions are now taking shape in reality. RIA is happening parallel to and in conjunction with SOA and it looks like many users will have interesting stories to tell later in the year.
At EclipseCon this week, the Eclipse Foundation announced that it is forming a new open source community project “to develop and promote open source runtime technology based on Equinox, a lightweight OSGi-based runtime.”
Mike Milinkovich, executive director of the Eclipse Foundation, told SearchSOA that this is important news for architects and developers working on service-oriented architecture (SOA) projects for three reasons:
- “OSGi itself and Equinox as its implementation has a service-oriented component to it. It is a technology that you use to pull together services in a runtime.
- “EclipseLink, which provides persistence to enterprise applications for storing either relational data or XML Schema supports the acronyms enterprise architects love like FDO [Feature Data Objects]. You can get implementations of that specification through EclipseLink.
- “It is part of the Eclipse Swordfish project, which is a full SOA runtime.”
When Swordfish was announced earlier this year, Anne Thomas Manes, research director for Burton Group Inc., said OSGi added “real value” and is a good fit for the Eclipse plug-in philosophy.
“There’s a lot of nice features to OSGi,” Manes told SearchSOA. “You deliver software in something called a bundle. As part of the bundle it identifies the manifest of all the things that are in there and also identifies the dependencies that this code has. Then the OSGi runtime can look at it and say in order to deploy this I have to get these things that are listed in the dependencies, and get those installed first. It’s a very clean and elegant way to package stuff up. The idea here is that you are going to package up services using OSGi.”
There is currently a discussion thread on TheServerSide.com regarding Equinox, EclipseLink, OSGi and its relation to the Java Community Process work on the Java Persistence API (JPA 2.0).
Rumor has it that the SOA market is on the brink of another wave of acquisitions. Oracle opened the year with its purchase of BEA Systems, but the move has yet to touch off much of anything in the way response. Possibly part of that is the deal took two major buyers off the market while Oracle ingests BEA.
Yet another part of that is the economy. Would-be buyers want to make sure they’re making smart acquisitions. What sort of revenue stream are they picking up and how much is that worth? First quarter financial results loom as a major influence in answering those two questions. That’s where the March Madness kicks in.
A niche player who can post a strong first quarter could position itself as the “gotta have it” item on the SOA market. You users out there might be thinking, “Big deal, this doesn’t really affect me.” Yet it does. If those vendors are hungry enough for quick revenue, they might be cutting some handsome deals over the next two weeks in order to pump up those balance sheets. Users might be able to land some best-of-breed technology at a discount and then have a large vendor step into place to provide ongoing support for that technology.
We know Iona Technologies is up for sale. Many larger vendors have data services and SOA testing gaps. Does HP look to flesh out its management story? Does SAP make a move into BPM? What can Tibco, Software AG and Progress Software do to push themselves over that $1 billion revenue mark? Are RIA, composite application and enterprise mashup technologies where the money is in the current market?
Don’t be surprised to see some clearance prices out there. The looming consolidation in the SOA space could create a buyer’s market.
Service-oriented architecture (SOA) expertise is still not available off-the-shelf.
That’s the reason Red Hat Inc. bought Amentra Inc., a integration services provider headquartered in Richmond, VA., which specializes in providing SOA knowledge transfer for its clients. In making this deal, Red Hat is betting that Amentra can provide the consulting services needed to support JBoss, the middleware company Red Hat acquired two years ago.
In a recent Q&A interview at JBoss World, Craig Muzilla, vice president of middleware business at Red Hat, talked about the pain points organizations run into when tackling SOA.
In an interview after the Amentra deal closed this week, Muzilla stressed how important SOA expertise is to the middleware market in general and JBoss in particular. He said companies making the transition from legacy mainframe or client/server to SOA often lack the expertise in-house to do the job.
“Amentra has a unique methodology focused around knowledge transfer,” he said. “Not only do they help design SOA and help the customer do some projects and implement project, but they also transfer that knowledge so the customer can be more self-sufficient.”
Bradley F. Shimmin, principal analyst of application infrastructure at Current Analysis LLC. agreed that knowledge transfer is one of the strengths Amentra adds to Red Hat and JBoss. Saying that this acquisition is “a perfect fit for Red Hat,” he noted that existing consulting services for JBoss had relied heavily on partnerships, and were not a match for the consulting services offered by the larger SOA vendors, such as IBM. The Amentra acquisition will begin to help close that gap.
Providing consulting in support of JBoss may be critical if Red Hat is too rearch its announced goal of capturing 50 percent of the enterprise middleware market by 2015.
In the blogsphere, Red Hat has received some criticism for its marketing of the JBoss products, which Muzilla sought to clear up earlier this week on Dana Blankenhorn’s ZDNet blog.
After the Amentra deal was announced, Larry Dignan, also blogging on ZDNet, wrote: “The deal, announced Thursday, gives Red Hat some foot soldiers to sell the company’s stack of software including JBoss. which has been a tough sell.”
Of course, Amentra is not on a par with something like IBM Global Services.
Shimmin notes that Amentra is based on the East Coast and that is where most of its clients are located, although it is doing work as far West as Chicago and Texas. The company is looking at expanding further West to the Pacific Coast. Plans to have any European or international operations seem to fall into the yet-to-be-determine category.
Marketers in the service-oriented architecture (SOA) world seem to be falling all over each other to make their new products Web 2.0 buzzword compliant.
Although Web 2.0 is a dubious term technically since there is no real Web 2.0. It is a clever catchall phrase for the more glitzy browser applications that emerged originally with wikis, and blogs, as well as Podcasts, which is another buzzword for downloadable digital audio files.
A chart of Google Trends data on Web searches indicates that Web 2.0 first came on the scene in mid-2004, when SOA was already flying high as a frequently searched topic. But after sliding under the radar for the next year, Web 2.0 took off like a rocket in late 2005 and surpassed SOA in the fourth quarter of 2006. Since then Web 2.0 has been the more popular term.
So it is perhaps not surprising that marketers are hyping their Web 2.0 capabilities in product announcements.
This week in announcing OpenLibertyJ , its open sourcing of Liberty Alliance security and privacy framework the major emphasis was on Web 2.0. SOA got only one mention in passing.
Asked why the big emphasis on Web 2.0, Brett McDowell, executive director, Liberty Alliance, said: “From my perspective service-oriented architecture is a concept that immediately resonates and gives you a vision of applications if you’re an enterprise architect. Web 2.0 gives you a vision of applications that are taking the Web by storm. What we wanted to use is the term that’s going to convey the correct expectation of what this framework is meant to enable.”
But that didn’t mean OpenLibertyJ had little or nothing to do with SOA.
“It absolutely enables the identity bus for SOA,” McDowell said. “But I think a broader audience understands the vision of Web 2.0.”
Jason Bloomberg, senior analyst for ZapThink LLC., was asked if this explanation was more about marketing or technology.Replying by email, Bloomberg wrote: “Technically correct or marketingese? Well, both. 100% marketingese with just enough truth mixed in :-).”
The Liberty Alliance is not alone in hitching a product wagon to the Web 2.0 star. Since 2006, Oracle Corp. has been talking about the convergence of the Java Enterprise Edition and Web 2.0 into something Thomas Kurian, Oracle’s senior vice president, called SOA 2.0.
That term does not appear to have caught on, as a request to Google Trends brought back this reply: “Your terms – SOA 2.0 – do not have enough search volume to show graphs.”
In 2007, Oracle began using the term Enterprise 2.0 for the Java, SOA and Web 2.0 convergence that is bringing wikis, blogs and social networks into the corporate world. Since first appearing on Google Trends charts in the fourth quarter of 2006, Enterprise 2.0 has been a hotter topic in Web searches than SOA 2.0. But when compared with SOA and Web 2.0, Enterprise 2.0 is still a flat line under their arcs.
If Oracle with its marketing muscle cannot get SOA 2.0 or Enterprise 2.0 off the ground, we may be stuck with Web 2.0, nebulous as it may be.
Discussing IBM’s SOA and Web 2.0 marketing strategy this week, Stephanie Martin, new worldwide lead for IBM Developer Relations, which includes more than 6 million coders who frequent the developerWorks Website, says she believes the two terms can play well together.
“They’re both very hot topics in the market right now,” Martin said. “In order to have the Web 2.0 experience, SOA is critical for designing and architecting these applications. That’s where I see the link between SOA and Web 2.0. Certainly they are not the same thing. SOA is the enabler of Web 2.0 but I do not see one replacing the other. We’re seeing our community’s interest in both those technologies growing consistently.”
So it appears SOA and Web 2.0 will have to co-exist as buzzwords, at least, until the next hot term is coined.