webMethods Weaves Rules Into Fabric
Type: Competitive Intelligence Report
Analyst: S. Willett
Report Date: December 7, 2005
Module: Application Infrastructure
ID: CIR21120 |
Current Perspective: Neutral
Vendor Importance: Moderate/High
Market Impact: Low/Moderate |
|
Summary
Event Summary
December 5, 2005 - SOA Software, a leading provider of comprehensive SOA and Web services management, security, and governance solutions, announced the acquisition of the X4ML Mainframe Web services platform from Merrill Lynch. The purchase of this technology provides SOA Software with a production proven, unique Mainframe Web services solution that complements SOA Software's existing products to create a complete SOA Infrastructure solution addressing B2B, legacy, and distributed applications.
Analytical Summary
• Current Perspective: Neutral on SOA Software’s new legacy Web Services product, as it does fill a definite hole in the market; however, it is only marginally related to the firm’s governance suite.
• Vendor Importance: Moderate to high to SOA, as this opens up a completely new market and revenue stream for the firm, although it may not drive sales of other products.
• Market Impact: Low to moderate on the Integration & Web Services market, as few in the SOA governance space will get into Web Services creation in this manner, although this product should raise alarm bells at IBM and other host integration firms about the need for “on the mainframe” Web services tools.
Target Markets
Global 2000, Global Carriers, Large Enterprises, Mobile Wireless Operators, Resellers/Channels, Systems Integrators, Third Party Implementers, Web Portals
Perspective
Current Perspective: Neutral
We are taking a neutral stance on SOA Software’s move to purchase the X4ML Mainframe Web Services technology from Merrill Lynch (SOA is privately held and did not disclose the purchase price), which had developed it for internal use. While there is probably a market for this Cobol/CICS oriented technology, it takes SOA far afield from its SOA governance market, where it has some differentiators.
SOA Software is one of several SOA governance or management vendors in the market. The company, however, is differentiated by its early bet on UDDI 3.0 as a means to organize and manage services, as well as discover them. It turns out that users do want a registry in the profoundly distributed world of SOAs, and UDDI 3.0 is really the only standard way to achieve this. The company uses UDDI 3.0 “TModels” to extend descriptions of services with policy-oriented rules and even security parameters. This is also unique, although it isn’t clear if TModels will be the way to extend services, as the larger companies have not yet endorsed this notion. In any case, SOA Software as well as Systinet would have a technology that would seem to be in high demand from ISVs and users.
Now it is complementing this with software that more easily enables Web services development out of traditional CICS/Cobol applications on the mainframe. It’s new Service Oriented Legacy Architecture (SOLA) product will open up a new market segment for SOA Software, and could conceivably attract mainframe shops to SOA’s larger suite. SOLA is unique in that is designed for CICS/Cobol programmers and allows exposure of key processes and information as a Web Service. This is opposed to “off the mainframe” solutions that basically grab mainframe data and then expose it as a Web Service on an intermediary J2EE or .Net platform.
The firm also says that SOA is differentiated from “on-the-mainframe” solutions that make use of Cobol copybooks, in that it encapsulates the entire Cobol/CICS process. The product has been working in an extremely large production environment, Merrill Lynch, which gives it some credence and maturity. It supports several transports, including MQ and HTTP, and includes an integrated test harness. SOLA will be integrated with SOA Software’s UDDI 3.0 registry in that new services will be automatically registered in the SOA Registry, and can be managed through the firm’s Service Manager governance module (proxies for availability, however, will not run on the mainframe at this time).
While there is probably a market for SOLA, it is only marginally related to SOA Software’s SOA registry and governance products. SOA Software is a small firm, and must now design marketing and sales programs around reaching a completely new market. This could be a distraction and create an inefficient business model for the firm.
The move also puts SOA into competition with a wide range of legacy integration players, notably IBM with its CICS Gateway product, as well as many smaller firms who have both “on the mainframe” and “off the mainframe” Web services solutions. IBM has, in fact, been slowly improving its CICS gateway product (see “IBM Gives SOAs a CICS,” December 1, 2004). Many others compete in this space. In general, “off-the-mainframe” solutions have gained favor because firms have made strategic decisions to put middleware and new programming on Java or .Net application servers. They may not welcome a Cobol/CICS product on these grounds, no matter what the productivity benefit is.
SOA should develop a realistic plan to market this technology, including to organizations outside the specialized Merrill Lynch type space (the company’s approximately 2,000 Cobol programmers put it in a very unique position). It should also renew investments in its core SOA governance products, including more management and synchronization options.
It is highly unlikely that other SOA governance or management firms will follow SOA Software down this path. However, those firms that build host integration software need to continue to ease programming with more automation and point and click Web Services support.
Positives and Concerns
Competitive Positives
• SOA Software purchases the X4ML Mainframe Web Services technology from Merrill Lynch and intends to market this as Service Oriented Legacy Architecture product by early next year. This opens up a new market segment for SOA Software, which is making its mark in SOA governance, registry, and management. It also has a B2B VPN product. The move could conceivably attract mainframe shops to SOA’s larger suite, allowing it to gain more traction against Systinet and others in the SOA governance space.
• The new SOLA product is unique in that is designed for CICS/Cobol programmers and allows exposure of key processes and information as a Web Service. This is done in an automated way, with a developed user interface. This is differentiated from the majority of tools that attempt to take the services off the mainframe and use Java application servers or another intermediary platform. For whatever reason, many programmers and organizations find that this “off-the-mainframe” method is difficult and creates a lot of performance overhead. So SOLA could be attractive to these users.
• As opposed to other “on-the-mainframe” solutions, the tool allows programmers to more directly encapsulate the CICS business process in a Web Service and doesn’t require the copying of Cobol copybooks, which the company argues doesn’t capture the actual context of the service. In general, this is an underserved market, as few are eager to cater to the Cobol/CICS programming market, which is in slow decline. In that sense, SOA Software has a differentiated offering.
• The product has been working in an extremely large production environment, Merrill Lynch, which will buoy confidence among corporate users as to its maturity. It supports several transports, including MQ and HTTP, and includes an integrated test harness. SOA is also bringing over a few key people from Merrill Lynch. The company has a few early access users now on the commercial product, which will be available in January 2006.
• SOLA will be integrated with SOA Software’s UDDI 3.0 registry. New services will be automatically registered in the SOA Registry, and can be managed through the firm’s Service Manager governance product. This will allow mainframe programmers to browse through the registry in order to find and consume services running elsewhere, which is a requirement in many shops. This could drive sales of the entire SOA Software Suite.
Competitive Concerns
• An “on-the-mainframe” Web Services product is only marginally related to SOA governance, which is the high-growth market SOA Software has been aiming for. This puts it in a totally different market, with different channels, selling methods, marketing programs, and expertise required from sales and technical staff. This could be a distraction and create an inefficient business model for SOA Software, which is a small firm.
• The move also puts SOA into competition with a wide range of legacy integration players, notably IBM with its CICS Gateway product, as well as many smaller firms who have both “on-the-mainframe” and “off-the-mainframe” Web services solutions. IBM has, in fact, been slowly improving its CICS gateway product in response to requests from Cobol/CICS users.
• Many firms have made strategic decisions to put middleware and new programming on Java or .Net application servers and may not welcome a Cobol/CICS product on these grounds, no matter what the productivity benefit is. This simply has to do with a long-term move away from proprietary Cobol applications. Also, since it originated as a user technology from Merrill Lynch, it is likely very well suited to its (admittedly very large) stable of programmers. It isn’t clear if the features in it are equally attractive to other firms with different application configurations in different verticals.
• SOA registries and governance remains a promising market, and it isn’t clear why the firm isn’t investing here. Most major middleware firms (most recently IBM) have committed to both a registry and repository, basically endorsing SOA Software’s direction. The OEM market in particular could be a large market for SOA Software. Users also have basically indicated that a registry and governance is needed once services begin proliferating.
Recommended Actions
Recommended Vendor Actions
• The firm should concentrate on its registry and governance line. In particular, synchronizing and replicating two different platforms and “proprietary” registries remains a problem that SOA Software should address. It is one of the few that seems to have a handle on how to manage disparate UDDI “Tmodels,” and it should expand its efforts there.
• The firm should also increase basic management functions for someone who needs to ensure services and the processes around them are up and running and performing well. Other performance information can be sent to systems management platforms. Some knowledge about SOA processes and their dependencies should be included in the service manager product, as it has a bearing on the performance of individual services.
• The firm should set up the appropriate channels, sales training, and marketing campaigns to reach the Cobol/CICS market with its product. It should also heavily promote its Service Registry and Service Manager to these users.
• The firm should “generalize” the SOLA product, so it widely appealing to users outside of Merrill Lynch and the financial services industry. It should set up points of differentiation with IBM’s CICS Gateway product.
Recommended Competitor Actions
• Competitors in the SOA Governance space, including Systinet, Infravio, Amberpoint, Fujitsu, Software AG, and others, should not necessarily enter the mainframe Web Services integration space at this juncture, as it would likely be a diversion.
• Firms with mainframe integration software, including IBM, TIBCO, Iona, Neon, and others, should continue to ease development of Web services from CICS/Cobol applications through better user interfaces and automated procedures. They should monitor SOA Software’s approach to see if it gains favor among Cobol/CICS programmers.
• In general, integration competitors need to address SOA governance, since SOA integration will involve managing and keeping track for services that were not developed on their platform. This could involve simply ensuring that runtime registries can be synched up to a third-party UDDI registry, all the way up to offering a full-fledged registry and repository of services, with some basic management functionality.
Recommended End User / Customer Actions
• End users with a significant amount of Cobol/CICS programming may want to consider this tool as a way to expose more of the information and processes as Web Services without setting up intermediary integration and middleware systems.
• End users will also want to evaluate more traditional host integration software from IBM, TIBCO, Neon, Iona, and others.
• In general, users interested in setting up SOAs, with Web Services as the key type of services, should consider a registry and SOA governance and management in their plans. This will become more of an issue as services proliferate and applications become more mission critical.
|
