Planning Assumption

2000 Forecast for the Application Integration Market

June 15, 2000

Mike Gilpin

Giga Position

Ever since the market for application integration solutions first began to emerge in 1997, the market has grown rapidly, first as enterprise resource planning (ERP), and later as customer relationship management (CRM) and e-business investments drove a significant ongoing requirement for integration of new initiatives with legacy or packaged applications. In the last year, the correlation to e-business has increased, to the extent that this now drives about 65 percent of the new uses of application integration solutions.

Reflecting these different uses of application integration solutions, the market has developed along parallel lines, leading to multiple market segments in which different vendors have established leadership depending on their special ability to dominate a particular segment. This combination of new uses, new vendors and growth in all the segments is driving rapid growth in the category as a whole, at what Giga predicts will be a rate of 57 percent growth for 2000 results over those in 1999. This will result in a market size for 2000 of $2.1 billion in products and closely related services, or of $1.3 billion in product license fee revenue alone.

Proof/Notes

Giga has been tracking this market for three years now, and it has reached a stage where it is possible to size the market with a reasonable degree of concrete data, since so many of the players are publicly traded, and for which it is possible to assess what share of their revenue comes from their application integration business.

Market Methodology

·     The vendors included in this analysis were IBM, BEA Systems, New Era of Networks (NEON), Mercator, Software Technologies Corp. (STC), TIBCO, Vitria, Active Software, Constellar, Viewlocity, OnDisplay/Oberon, SOPRA, Software AG, Information Builders, NEON Systems, SAGA Software, Level8, CommerceQuest, Sun/Forté Fusion, and Microsoft (which will be a bigger player in 2001, but BizTalk revenue won't start accruing significantly until later in 2000).

·     To be included, the vendor or the line of business had to be primarily identified as an application integration solution, including Enterprise Application Integration (EAI), Internet Application Integration (IAI), and customer-centric or "front-office" integration solutions. For this calculation we chose to exclude revenue from webMethods, Extricity, Ariba, and other major players that are primarily positioned in the business-to-business e-commerce solution market, although the lines are fuzzy here. With the acquisition of Active Software by webMethods, as other similar changes occur it may be necessary next year to consider these markets together as one.

·     For those vendors included, we considered all the revenue from the application integration line of business, both product and service. For those wishing to break out just product revenue, use an average for the group of 60 percent of revenue derived from products, making the 2000 product revenue market size $1.3 billion. Note however that our definition of service revenue in this context counts mainly services tied directly to the implementation of the application integration solution, not including more general systems integration. Although for application integration projects, we typically see a 50/50 split between product and service expenditures, the vendors tracked here average about 60 percent in product revenue because several partner with systems integrators to deliver a significant proportion of their service engagements.

·     In assessing 2000 revenue, we compared the first quarter of 2000 to the fourth quarter of 1999 for a quarter-on-quarter growth rate, which we then applied to the whole year 2000, in most cases. In a couple of cases, we factored in other considerations to make adjustments to these projected growth rates, such as recent or impending acquisitions, or lines of business which are likely to be shed. We also looked at all four quarters of 1999 revenue to do a sanity check on the numbers and spot any seasonal variations, but for most vendors, the growth from the first quarter of 1999 through the first quarter of 2000 has occurred at very consistent rates. There were a few vendors, though, that hit a clear "hockey stick" during 1999, including Vitria, Active Software, TIBCO and Level8, and we took this into account in making a realistic assessment of their 2000 growth rates.

·     We estimated IBM's revenue by tapping a number of industry sources, since IBM does not break out MQSeries revenue for financial reporting.

Vendor Analysis

The results of our research are tabulated (see Table 1) and summarized graphically (see Figures 1 and 2), surveying these results:

·     For 1999, we estimate this market reached $1.3 billion in revenue.

·     In 1999, the market leaders were IBM with a 34 percent share, New Era of Networks and Mercator each with a 9 percent share, and TIBCO and SOPRA with a 7 percent share. IBM's numbers appear somewhat inflated, because we are including all MQSeries revenue in the number for IBM, although one might argue that some of this revenue comes from uses of MQ that predate usage in this market. However, today, virtually all Giga clients using MQ are using it to do application integration, so we have chosen to count it all. Take this into account in comparing results to other vendors without a significant position in business-quality messaging technology. Also, some other vendors bundle their messaging technology and thus don’t generate explicit revenue from it, like Microsoft with MSMQ.

·     For 2000, we forecast the market will reach $2.1 billion in revenue, a growth rate year-on-year of 57 percent. The market is growing faster than last year (rather than the slowing in growth that we had earlier anticipated), primarily through the entry of new vendors into the market and the expansion of the market into adjacent business-to-business (B2B) e-commerce solutions. Although growth rates will still probably decline somewhat in future years, we estimate this market will reach $5.5 billion in size by 2003. Whereas today about 65 percent of this market is driven by e-business, by 2003, that will have grown to at least 80 percent, leaving only 20 percent of the market (about $1.1 billion) for non-e-business uses like internal integration of ERP and legacy applications not connected with e-business. Thus, e-business integration will average a compounded annual growth rate (CAGR) of about 50 percent during this period, while non-e-business integration will grow at a much more modest CAGR of about 15 percent, reflecting a mature market.

·     We predict that the market leaders by the end of 2000 (based on 2000 results) will be IBM (29 percent share), New Era of Networks (10 percent share), and TIBCO and Mercator each with about 9 percent share. The most significant increase in share among these vendors compared to 1999 results will be TIBCO, which is getting a lot of traction from various e-business and other "front-office" related integration efforts.

·     The next group of vendors is predicted to have share in 2000 ranging from 3 percent to 6 percent, led by Vitria and SOPRA, followed by STC, Viewlocity, Information Builders and Active Software (now to become part of webMethods, but considered separately for this analysis, since the acquisition hasn't closed yet) and Level8.

·     A third group of vendors that we predict will have share in the range from 2 percent to 3 percent includes BEA Systems, OnDisplay/Oberon (now to become part of Vignette), NEON Systems and Forté Fusion. Depending on how one allocates revenue from the WebLogic product line, one could consider BEA to be a bigger player in this market, but for this analysis, we included only BEA's eLink revenue, which today is based on the Tuxedo line of business.

·     Another way of viewing this market is to factor front-office integration from back-office integration (see Figure 3). Looking at those vendors dominating the front-office category (associated mostly with CRM and/or e-business implementations), the leaders for 2000 will be TIBCO, Vitria and Active Software.

Table 1: Market Share of Application Integration Vendors

Source: Giga Information Group

Note that a difference of one percentage point one way or the other should be considered within the range of likely outcomes bracketed around the predicted figures in Table 1. A "+" designation indicates somewhat more potential to the upside of that figure.

Market Analysis

Although this market superficially resembles the application server market in the way a few larger players dominate the numbers, followed by a large group of vendors with lesser results, the market dynamics are not the same, for two reasons:

1.        Whereas one Enterprise JavaBean (EJB) application server has a reasonable shot at replacing another, application integration solutions are, by and large, not interchangeable. Today, they are mostly implemented on much more proprietary technology, and once a solution is implemented, it is more likely to remain a strategic component (i.e. "locked in") over time.

2.        There is also a great deal more identifiable segmentation in the intended purpose of various application integration solutions, by application type, vertical industry and correlation to front or back office.

These two factors act together to keep this market from becoming one large ecosystem, instead creating a number of smaller "micro-ecosystems" in which different groups of vendors are able to have local dominance. Figure 3 is the latest market segmentation chart, as presented at GigaWorld IT Forum 2000.

The overall growth rate of the application integration solution market is not as great as that of the application server market (57 percent vs. 178 percent). Although both markets are being driven substantially by the same external growth factors (including e-business, CRM and mobile applications), the application server market is newer and hence in a hotter growth phase, and is also on a trajectory to become at least twice as large in the long run. The key difference driving this higher trajectory is the fact that application servers are becoming fundamental infrastructure to most emerging applications in these emerging categories, whereas only the larger and more complex integration scenarios are tending to drive purchase of application integration solutions.

As the market continues to evolve, there are a number of forces operating to change this market dynamic:

·     As discussed in earlier research (see Planning Assumption, Emerging Standards-Based Application Integration Solutions, Mike Gilpin), the application integration and application server categories are merging, becoming less distinct separate markets, driven by many factors, the greatest of which is the desire of customers to obtain one more unified solution set from a smaller number of vendors (ideally one) across both categories.

·     Application integration solutions are becoming increasingly standards-based, making them more interchangeable. At the same time as this creates opportunities for smaller players to add value around solutions from larger vendors, it also moves the industry inexorably toward being more like the application server market, with a few dominant large vendors and a larger number of niche players.

As this transition occurs, for the vendors in the "niche" group, it will become increasingly important that they have a clear understanding of what their niche is, and a technology, marketing, and sales strategy to match that niche, just like in the application server market (or any more mature market). However, because of the increasing correlation between these solutions, some of the effective niche strategies may be the same, especially where the niche has a significant requirement for application integration. Such niches include mobile applications (which need complex synchronization capability across a number of dimensions with their nonmobile counterparts), as well as specific lists of applications that are applicable within each of the various vertical markets a solution might aim to focus upon. So, just as Active Software and Vitria are strong in telecommunications firms today, they are likely to maintain that strength going forward if they are able to keep their solutions relevant as the nature of the technology infrastructure continues to evolve.

Another scenario for some of these vendors not in the top group is to be acquired by a much larger company that has a significant sales channel already selling to the right audience. In this category, these may be application server vendors, or application vendors in the category in which the integration solution is most relevant. The example of webMethods and Active Software also points to another fertile ground for acquisition (see IdeaByte, Internet Application Integration Category Heats Up With Acquisitions of Active Software and OnDisplay, Mike Gilpin), unifying application integration with a relevant e-business solution domain.

Alternative View

The next most likely scenario is just a more pessimistic one, with the pessimism arising along one or both of two dimensions:

·     Less market growth because of increasing disillusionment with the category: Although Giga still believes in the fundamental value of application integration solutions, we continue to hear persistent stories of some implementations that do not achieve the anticipated value, or which disappoint in some other way. While these circumstances are not unusual for new products in an emerging market, some of the products are now into their second or even third generation, and customers have, in some cases, lost patience with their suppliers. Today, that means switching to another supplier, but in the future, it could mean finding other ways of meeting the need for application integration, hurting the growth of the category.

·     Even more rapid consolidation than is built into current assumptions could occur if smaller players begin to experience growth too slow to keep forward momentum, especially if this were to occur at a time when equity markets are off their peak. Extremely rapid consolidation, although it would still result in revenue growth for the acquiring vendors, could result in slower growth for the category as whole, especially to the extent it represents the failure of the smaller vendors being acquired. To date most acquisitions in this market have reflected success rather than failure, but that could change.

One of the reasons for disillusionment with this category of solutions is their continued reliance upon proprietary technology. The trend to more standards-based approaches discussed above could rescue this category just as it most needs it. Customers' willingness to invest in creating software assets expressed in a proprietary way (like the business rules in a message broker) will decline as alternative, standards-based ways of achieving the same objective become available, which should drive a more rapid transition to the new solutions than might now be evident. There are other important gating factors that could act to limit the growth of this market:

·     Even with tools, the effort and duration of application integration projects can be large, and large projects are more likely to fail. Staged implementations can help, but if failures occur and are blamed on vendors (always a convenient if sometimes well-deserved scapegoat), it could contribute to further disillusionment with the category.

·     As with all e-business projects, these integration efforts encounter limited supplies of the skilled resources required to understand integration requirements and to effectively grapple with integration toolsets to achieve the desired result.

·     Limited solutions from most vendors today for managing the profusion of metadata that arise in the context of multiple application integration efforts contribute to a reduction in the opportunity to harvest or otherwise reuse the results of one integration effort in another. Until repository-based solutions for managing this metadata, and the tools support them, are more widespread, this will be a continuing problem that will sap the strength of integration efforts over longer periods of time.

·     Event-driven application integration solutions, while useful, are not a panacea, and will not displace alternatives like extract, transform and load (ETL) solutions, at least not where those continue to be well suited to requirements to move large quantities of bulk data. These two ways of interfacing systems are just points on a continuum of solutions for moving data and events in the context of a business process.

Findings & Recommendations

Today, clients should consider these industry dynamics in the context of the overall set of weighted evaluation criteria discussed in Planning Assumption, How to Select an Enterprise Application Integration Solution, Mike Gilpin. The weighting and criteria along the seven dimensions will not only be specific to each company's context, but will also be strongly influenced by the particular vertical and application domains in which integration is to be performed. These company-specific criteria should be the main factor driving selection, although vendors' position in the market should also influence which vendors make the cut going to the short list. However, at this stage of market evolution, for specific integration scenarios, it can still be very much the best choice to go with a vendor that has the best fit to purpose, even if that vendor has only a 2 percent or 3 percent share of this market, as long as it has sufficient size overall to ensure continued product evolution and good product support.

For clients assembling e-business infrastructure, it is also appropriate to focus the selection of application integration solutions into alignment with the same infrastructure standards for communications, components, connectivity and data, like XML. This will naturally drive selection of solutions that are either from the same vendors that provide one's strategic application servers, or else are sufficiently aligned so as to interoperate with those servers smoothly.

When choosing an application integration solution that aligns with one's application servers of choice, give preference to those vendors that show leadership in the niches which are most closely aligned with one's vertical market alignment. This will continue to be important, since skills and knowledge of integrating applications in that vertical domain will tend to reside with those vendors, even as the underlying infrastructure technologies continue to evolve. This is just a special case of choosing systems integrator talent from those firms that know your business; in this case, that talent is embodied not only in consultants, but also in the product developers that build integration components that are closest to the application domain, such as adapters or connectors.

References

Planning Assumption, 2000 Forecast for the EJB Application Server Market, Mike Gilpin and Carl Zetie

Planning Assumption, Introducing an Internet Application Integration (IAI) Architecture Framework for E-Business, Mike Gilpin

Planning Assumption, Emerging Standards-Based Application Integration Solutions, Mike Gilpin

Planning Assumption, How to Select an Enterprise Application Integration Solution, Mike Gilpin

IdeaByte, Internet Application Integration Category Heats Up With Acquisitions of Active Software and OnDisplay, Mike Gilpin

IdeaByte, TSI to Acquire Novera: Another Sign of Application Servers and Integration Solutions Merging, Mike Gilpin

IdeaByte, STC eGate New and Improved but Not a Threat to NEON, Mike Gilpin

IdeaByte, Oracle Integration Server: Promising Vision, Limited Proof as Yet, Mike Gilpin

IdeaByte, NEON Integrates NEONImpact Message Broker (From CAI Acquisition) With MQSeries Integrator: Integrating the Integration Solution, Mike Gilpin

IdeaByte, Will Application Server Vendors Take Over the Application Integration Market? Mike Gilpin

© 2000 Giga Information Group