TIBCO Prepares for SOAs and ESBs with Matrix
Type: Competitive Intelligence Report
Analyst: S. Willett
Report Date: June 30, 2005
Module: Application Infrastructure
ID: CIR14076 |
Current Perspective: Neutral
Vendor Importance: High
Market Impact: High |
|
Summary
Event Summary
June 27, 2005 -- TIBCO, a business integration and process management software company that enables real-time business, announced Project “Matrix,” a deployment platform for service-oriented architecture (SOA) that enables composite applications to be developed, deployed, and managed as distributed, standards-based services. As part of Project Matrix, TIBCO also announced support for the Java Specification Request (JSR) 208, the Java business integration (JBI) specification. The first product shipped as part of Project Matrix will include a JBI service container, which will be available to early access customers in H1 2006.
Analytical Summary
• Current Perspective: Neutral on TIBCO’s JBI “Matrix” product, because the firm has some good ideas, but its attempt to attract the JBI SOA integration market and challenge ESBs is too undeveloped at this stage.
• Vendor Importance: High to TIBCO, since the company needs to form a response to the next wave of changes in the integration market and this product seems to be a leading element of that response.
• Market Impact: High on the integration and Web services markets, because all competitors need to adjust their strategies and users need to study their proposed solutions.
Target Markets
B2B Communities, End Users, Global 2000, Resellers/Channels, Systems Integrators, Third Party Implementers, Web Portals
Top
Perspective
Current Perspective: Neutral
We are taking a neutral stance on TIBCO’s plans for a JBI-based product focusing on SOAs, dubbed “Matrix.” The company is rightly preparing for SOAs with a discrete product, but it has not filled in many details (including pricing) concerning the product, which will not ship until early in 2006.
The notion that SOAs will require (or benefit from) new types of integration products has been floated by vendors, mainly smaller ones, for the last several years. These vendors, which call their products “ESBs” (enterprise service buses), have sought to map out features that include intrinsic assumptions about services end points, some services management, a lightweight broker to guarantee message delivery, transformation, WSDL and Web services creation, and a process tool. They have talked about upending traditional EAI products with lower costs and lightweight, distributed processing. Now that SOAs are being implemented in earnest and services are proliferating, the need for an ESB is becoming more obvious. The proliferation of WS-* standards will only intensify this need (see “The Time Has Come for ESBs” February 14, 2005).
Thus, bigger players, notably BEA, have jumped into the game and Sun will follow suit as it digests SeeBeyond and builds up its JBI ESB product. Some larger players, notably IBM and TIBCO, have dismissed the notion of an ESB product class for obvious reasons (they have a large stack to protect!). However, ignoring trends is never a good strategy and TIBCO is now dealing with ESBs in its own way, which could bring opportunities for the firm, not just threats.
Matrix, TIBCO’s answer to SOAs and ESBs, includes enough TIBCO infrastructure to make SOAs safe for large, risk-adverse IT organizations -- its core constituency. The product, also dubbed “JBI-Service Delivery Platform,” builds on JBI’s barebones framework, with its own messaging, which will provide the “eight styles” of messaging, including pub/sub, queuing, multicast, and broadcast. This, combined with other technologies, will enable transactions. It will also include XSLT transformation capabilities, another specialty of TIBCO. The messaging, transaction, and transformation capabilities will allow mission-critical, event-driven, and transactional applications to exist in an SOA. TIBCO is also a JBI .Net container that will allow the management and control of service endpoints on a Microsoft platform. This overcomes a principal objection of JBI by vendors and users, and a weakness of most ESB products on the market. Another value-add is testing, monitoring, and management of the overall platform. These features will make SOA Integration more acceptable to large corporate clients with conservative IT departments.
TIBCO is adding features for building interfaces, using its interface building tool, and adding AJAX support for rich Java user interfaces. TIBCO has a host of other technology to bring to bear on Matrix that will allow customers to build composite applications on top of Matrix. This includes process tools and a full BPM suite, adapters, a portal, interface building tools, and, of course, BAM and B2B capabilities.
TIBCO, however, has been vague on many details regarding Matrix, (including pricing), which will not be available until Q1 2006. The firm needs to fill in more details on the product, particularly how it will handle transactions, what role its UDDI directory will play, how the .Net container will work, and what WS-* standards will be supported and how. The firm should explore UDDI and a registry as the organizing force around controlling and managing services, including those on .Net and other platforms. The firm should include a BPEL process tool in Matrix for the simple orchestration of integrations. It should also include better security features (both message and user), perhaps through a partnership, since this addresses the biggest concern of large users when it comes to SOAs. The firm’s support for composite applications on top of Matrix is a bit scattered. It should consider a separate composite application product that includes higher-level BPM, BAM, interface building, and portal functionality. This will require a reconfiguration of the product suite, including the iProcess products.
TIBCO’s moves will likely cause competitors, particularly IBM, to react in some way. IBM is the only major vendor that has not weighed in with an ESB or SOA integration type of product. While not all competitors will aim for TIBCO’s high-end space, there will be better management consoles on ESBs or SOA integration offerings. Users will also react to TIBCO’s plans, with more taking action in building SOAs and looking for discrete products to manage them.
Top
Positives and Concerns
Competitive Positives
• TIBCO announced that it will build a JBI (Java Business Integration) product dubbed “Matrix” that will be aimed at IT environments that have, or will have, a service-oriented architecture (SOA) in place. This is TIBCO’s answer to ESB products, and it includes a host of features for distributed networks, embodies standards in the JBI and WS-I space, is “service aware,” and forms a solid foundation for building composite applications on top of these SOAs. The firm is thinking ahead and has a credible answer to ESBs that will not change its fundamental high-end selling model. It is bound to appeal to its high-end customers that are worried about the security, performance, and availability issues associated with an SOA. In essence, TIBCO is making SOAs safe for large, risk-adverse IT organizations -- its core constituency.
• Matrix, alternately called the JBI-based Service Delivery Platform, builds on JBI’s barebones framework, which basically describes how different integration components interact with each other. It requires a JVM, but not necessarily a full-fledged J2EE application server. TIBCO is adding its own messaging layer on top of JBI’s “Normalized Message Router,” which will provide the “eight styles” of messaging, including pub/sub, queuing, multicast, and broadcast. This, combined with other technologies, will enable transactions. It will also include XSLT transformation capabilities, another specialty of TIBCO. The messaging, transaction, and transformation capabilities will allow mission-critical, event-driven, and transactional applications to exist in an SOA.
• TIBCO will add a .Net container that will allow management and control of service endpoints on a Microsoft platform. This overcomes a principal objection of JBI by vendors and users and a weakness of most ESB products on the market. Another value-add is testing, monitoring, and management of the overall platform, which TIBCO plans, partially through an UDDI directory. These features will make SOA integration more acceptable to large corporate clients with conservative IT departments.
• TIBCO is adding features for building interfaces, using its interface building tool and the new AJAX standard for building rich Java-based user interfaces for administrators and “configurators.” This, combined with its base application server capability, will allow users to build business logic and actual applications on top of this SOA.
• In fact, TIBCO has a host of technology to bring to bear on Matrix for building composite applications on top of an SOA. There are process tools and a full BPM suite, adapters, a portal, interface building tools, and, of course, BAM and B2B capabilities.
Competitive Concerns
• TIBCO is vague on many details on Matrix, which will not be available until Q1 2006. For example, how transactions will be handled is unclear (there are a number of WS-* standards in this space, as well as TIBCO technology), and it is unclear the exact role UDDI will have in the management of a network. The messaging seems to an amalgamation of existing TIBCO technology and it is unclear just how much it borrows from TIBCO’s existing product. The details of the firm’s .Net container were also not explained explicitly. The extent to which various WS-* standards will be supported is unclear, as some of them can theoretically be used in lieu of “proprietary” TIBCO technologies. Users will want more details on these elements before making buying decisions.
• TIBCO has not settled on pricing. Vendors such as BEA are staking a differentiator on lower pricing for an ESB than traditional integration technologies. This is a solution that, like it or not, will be compared to ESBs-branded products.
• The solution does not appear to include a process tool, a likely requirement for multi-step integrations and higher-level composite applications.
• In fact, the technologies for building composite applications seem to be diffused throughout several TIBCO products. Purchasing all TIBCO products for composite applications can be an expensive proposition.
• TIBCO is also relatively silent about security. Although it falls somewhat out of TIBCO’s realm, this is a major issue with SOAs among the type of clients the company is targeting.
Top
Recommended Actions
Recommended Vendor Actions
• TIBCO needs to fill in more details on Matrix, particularly how it will handle transactions, what role its UDDI directory will play, how the .Net container will work, and what WS-* standards will be supported and how. For the time being, the firm would be wise to combine WS-* with its own functionality, rather than relying on WS-* standards to handle transactions, messaging, eventing, etc.
• The company should explore UDDI and a registry as the organizing force around controlling and managing services, including those on .Net and other platforms. It may even want to make an acquisition in this space. It should front this with a graphics-laden management console that will be attractive to large enterprise sites.
• TIBCO is obviously aiming for a “high-end ESB,” which fits nicely with the firm’s current customer base and its selling model. Pricing should reflect this, and TIBCO should come in significantly higher than ESBs on the market and somewhat lower than EAI implementations.
• The firm should include a BPEL process tool in Matrix for the simple orchestration of integrations. It should also include better security features (both message and user), perhaps through a partnership, since this addresses the biggest concern of large users when it comes to SOAs.
• TIBCO should consider a separate composite application product that includes higher-level BPM, BAM, interface building, and portal functionality. This will require a reconfiguration of the product suite, including the iProcess products.
Recommended Competitor Actions
• Competitors in the integration space that do not have an ESB or a discrete product for SOAs should consider one.
• Competitors with ESB products should consider a management console, as well as a UDDI directory. In addition, they should consider an ebXML registry to manage and control SOAs.
• Competitors building products on JBI (and even those that are not) need to build in more features for controlling and managing services on .Net.
• Competitors should not necessarily consider the type of advanced messaging that TIBCO has in mind for Matrix, as that will likely be relegated to a high-end market.
Recommended End User / Customer Actions
• Users that are building SOAs should consider an integration product that takes advantage of an SOA’s unique characteristics, including re-usability, standards-based formats and methods, and distributed processing. In many cases, this will mean an ESB, which has the potential to be lower-cost.
• Users should consider a composite application environment as an accompaniment to an SOA, as this will allow for more flexible applications that draw on the information and events contained in “siloed” applications and databases. They should develop governance of these processes.
• Users that are building out SOAs should be concerned with governance of these services, including security and performance issues. They should seek out products that solve these issues designed specifically for SOAs.
|
