IBM Accelerates SOA Adoptions Through DataPower Purchase
Type: Competitive Intelligence Report
Analyst: S. Willett
Report Date: October 19, 2005
Module: Application Infrastructure
ID: CIR14735 |
Current Perspective: Positive
Vendor Importance: Moderate
Market Impact: Moderate/High |
|
Summary
Event Summary
October 18, 2005 – IBM announced it has acquired DataPower, a Cambridge, MA-based, privately held provider of products that help improve security and speed the processing of computer transactions. Financial details were not disclosed.
Analytical Summary
• Current Perspective: Positive on IBM’s purchase of DataPower, as it gives IBM hardware acceleration for its integration and “SOA” products, giving it an edge over competitors and allowing it to compete with Cisco and others who will likely enter this space.
• Vendor Importance: Moderate to IBM, as this type of message-level hardware acceleration is necessary to compete with Cisco and to attract top customers who need to deal with heavy XML transformations in an SOA environment.
• Market Impact: Moderate to high on the integration and Web Services market, as the move will cause integration competitors to offer their own hardware acceleration or at least to partner. It also increases the likelihood that more network equipment providers will enter this space.
Target Markets
B2B Communities, End Users, Global 2000, Resellers/Channels, Systems Integrators, Third Party Implementers, Web Portals
Perspective
Current Perspective: Positive
We are taking a positive stance on IBM’s purchase of privately held DataPower for an undisclosed sum, as IBM gains a product that it is calling an “SOA Appliance,” which performs XML acceleration as well as acceleration of security and other message-based processes. While a software division’s purchase of a hardware company seems a little odd, it does helps the firm compete against Cisco and other anticipated entries into the middleware space by erstwhile hardware firms.
In June, Cisco shook up the middleware space with its AON announcement (see “Cisco Disrupts Middleware Market With AON,” June 21, 2005). The product provides middleware, especially message-based middleware, which is tightly linked to hardware blades that can selectively accelerate processes and messages and give priorities to some of these operations. It also entered at a much lower price point than EAI vendors. While Cisco is missing many middleware elements that competitors have, most notably a process tool, this type of hardware/software combo is ideal for moving large volumes of messages in certain vertical applications. Now IBM sees benefits for hardware acceleration in the SOA space, where processing XML-related functions (particularly transformations and security operations) has proven to be onerous and process-intensive. IBM already has management software in Tivoli that can handle some SOA management and availability options in its integration and application server. And, of course, it has general-purpose hardware server to throw at the problem. The DataPower SOA Appliance will add another option, giving IBM the benefit of checking any advances Cisco or others (e.g., Nortel, Lucent) might make in middleware.
Specifically, IBM gains three products, one of which does straight-out XML and XSLT acceleration, another that offloads security processing, including security associated with Web Services (e.g., XML encryption, access control, etc.), and a higher-end offering that can accelerate any type of messaging, including non-XML message types found in WebSphere MQ and Message Broker. Once this is integrated into WebSphere and Tivoli consoles, it gives IBM an advantage over buying these products separately. The products have a software interface that allows users to put in rules around prioritization of messages by content, originator, or other metrics. This roughly matches Cisco’s capability in AON and could allow for better availability of certain operations in an SOA environment. IBM has big plans for the future of this technology – it plans acceleration of J2EE applications and DB2 operations. It also plans integrated administration with WebSphere, although the exact plans for a common console(s) have not been set.
On the negative side, this will lessen DataPower’s position as an independent product – though this is not a huge matter in the IBM scheme of things (DataPower had about 50 customers). However, if customers see it as a WebSphere-only device, they may turn to Cisco or others that support other application servers and message brokers more neutrally. Also, at this point, there are still two different interfaces for DataPower and WebSphere (and Tivoli), meaning there is little differentiation over integration players who simply partner with a hardware appliance vendor. IBM should hasten to add some common consoles and sell it as a bundle. DataPower’s price points are rather high compared to Cisco’s AON product (the higher-end product is $75,000, while AON starts at $40,000). IBM should match this pricing, for comparable functionality. To date, the XML processing market has been rather small and niche focused. IBM needs to both sell this as a general purpose SOA device, and target specific high-volume niches in select verticals.
Now that both IBM and Cisco have moved into middleware acceleration, competitors will likely respond. Expect more partnerships or even acquisitions by integration players. Also expect some network equipment providers to make some moves into this market, again through partnerships.
Positives and Concerns
Competitive Positives
• IBM purchases privately held DataPower for an undisclosed sum. The firm, which has 72 employees, has hardware appliances that accelerate the processing of XML, security, and other message-based integration operations. While hardware acceleration has been around for a while, a combo product could be a differentiator for IBM in the emerging “SOA Integration” market, where XML and related processing will become a user problem due to issues with XML and related security procedures. It also, notably, is a powerful alternative to the Cisco AON integration product, which is differentiated by hardware acceleration and prioritization. It is a pre-emptive move against other network equipment providers who will likely enter the middleware market in some respect in the future.
• DataPower has three products, one of which does straight-out XML and XSLT acceleration, another that offloads security processing, including security associated with Web Services (e.g., XML encryption, access control, etc.), and a higher-end offering that can accelerate any type of messaging, including non-XML message types found in WebSphere MQ and Message Broker. This can speed up SOA environments and offer an advantage over mixing hardware and software solutions from different vendors.
• The products have a software interface that allows users to prioritize messages by content, originator, or other metrics. Users can apply rules that can have specific effects on business use cases (e.g., giving priority to certain users, or transactions over a certain amount). This roughly matches Cisco’s capability in AON and could allow for better availability of certain operations in an SOA environment. Since SOA software management solutions are not foolproof, this gives an added layer of protection for mission-critical Web Services that are just now being employed.
• IBM has big plans for the future of this technology. It plans acceleration of J2EE applications and DB2 operations, as well as integrated administration with WebSphere, although the exact plans for a common console(s) have not been set. It will also explore blade- or even chip-level integration of these capabilities at some point. The integration administration, in particular, could save time and money over approaches where two different products are combined, or where one product or interface (e.g., Cisco AON) is used for some level of acceleration, and another set of products (e.g., TIBCO or WebSphere ESB) is used for other integration functions.
• IBM gains a modest revenue stream and 50 enterprise customers.
Competitive Concerns
• This will probably lessen DataPower’s position as an independent product, as IBM clearly is setting its sights on integrating it tightly with WebSphere. Users that don’t have WebSphere may feel left out of future plans, even if it does support generic XML processing functions. DataPower also had some relationships with other hardware providers, including Nortel, and it is unclear how those will proceed. While IBM has said it will maintain its AON partnership, it is clear that this solution will at least partially compete with it.
• To date, the XML processing market has been rather small and niche-focused, although the advent of SOAs and the heavy requirements of XML transformations and security may change that.
• The price points are rather high compared to Cisco’s AON product (the higher-end product is $75,000, while AON starts at $40,000).
• In the short-term, there are still two different interfaces for DataPower and WebSphere (and Tivoli), meaning there is little differentiation over integration players who simply partner with a hardware appliance vendor.
• Hardware-based acceleration and prioritization could eat into sales of software-only products that perform some of the same functions (e.g., Tivoli Composite Application Manager), and in general could serve to commoditize software with its low price points. While this is clearly in Cisco’s interest, it is not generally in IBM’s interest (at least not in the WebSphere group).
Recommended Actions
Recommended Vendor Actions
• Since IBM’s main differentiator will be a common WebSphere interface, it should hasten to build this. It should be an Eclipse-based interface with some common links to WebSphere and Tivoli administration consoles.
• The company should make it easier for users to formulate coherent rules and policies for prioritization and other matters (especially when traffic might flow across multiple appliances in a multi-threaded environment). It should also give guidance as to how this prioritization works with (or doesn’t work with) traditional software-based queuing and prioritization in the WebSphere MQ product.
• IBM should consider bundled pricing, with both the application server and/or the Web Services Management product. It should match Cisco’s AON pricing.
• The company should give rather explicit use cases that illustrate why this type of appliance can achieve advantages that can’t be achieved by software alone, using generalized hardware.
• The company should pick out a few target markets for marketing, likely in the telecom, financial services, and retail spaces.
Recommended Competitor Actions
• Competitors in the integration space should consider partnering with a hardware accelerator product in order to offer the same type of availability that IBM will offer in certain niche cases.
• Competitors in the application infrastructure and integration spaces should keep in mind that it is highly likely that other network equipment providers may be looking to partner in order to match Cisco and IBM in this middleware space. They should consider such partnerships.
• Cisco should continue to ensure that it can accelerate WebSphere MQ and Message Broker messages and should make this as seamless as possible. It should move to standards-based software (Eclipse, XSLT, WS-Security, etc.) whenever possible, and should support other messaging and integration interfaces.
Recommended End User / Customer Actions
• End users should consider SOA hardware acceleration in cases where performance is a problem and where generalized processing power and/or software-based management is not efficient enough.
• In general, end users should realize that XML transformation (XSLT, XPath, XSD) and WS-Security procedures will be process-intensive and could be a bottleneck. They should run pilots with heavy loads before rolling out mission-critical applications.
• End users should keep in mind that software-only solutions could solve some of the problems with Web Services performance, and can also identify the sources of problems through monitoring.
.
|
