RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
2008/04/07 00:00:00

Service-oriented architecture will destroy the ERP market?

Platforms implementing service-oriented architecture (SOA) are increasingly distributed. While they are used to integrate heterogeneous information systems, providing vendors with additional competitive advantages. But in the future, SOA is able to seriously change the market for ERP systems, making it similar to the market for OpenSource solutions, in which integrators who collect "distributions" of ERP systems from standardized services developed by many independent manufacturers will play a leading role.

Today, service-oriented architecture is unambiguously recognized by specialists as the "engine of progress" of corporate information systems for the next 5-10 years. Gartner predicts that by 2008, SOA will become a prevailing development practice, software defeating the dominant 40-year monolithic architecture. According to the company's analysts, the connection of SOA and open standards on which web services are based will bring this "tandem" to the forefront of technology development and in 2008, more than 75% of all development-related projects ON will be based on this new technology. According to a recent survey of IT directors conducted in the United States by magazines CIO and Computerworld, more than 58% of respondents have already implemented SOA or are going to do so. Of those involved in the transition to a service-oriented architecture, 44% will use it to integrate internal applications, 28% - to deliver services directly to customers, 21% - to communicate with external partner applications, the rest are going to work in all three of these areas at the same time. As a reason for the transition to SOA, 77% of respondents indicated an increase in information infrastructure flexibility, 61% - a reduction in the cost of integrating existing applications, 49% - a reduction in the cost of developing new applications and 47% - faster development of new applications.

SOA mania

Currently, in addition to market pioneers, IBM (WebSphere) and BEA Systems (AquaLogic), major players such as Oracle (Fusion Middleware), SAP (NetWeaver) and Microsoft (.NET) are promoting their SOA platforms. "",  according to "a magic square" of Gartner, Fujitsu (Interstage Suite), Novell (exteNd), Progress Software/Sonic Software (OpenEdge and SOA Suite), Software AG (Enterprise Information Integrator), Sun Microsystems (Java Business Integration), Tibco Software (Tibco BusinessWorks) and webMethods (webMethods Fabric) also participate in a game.

Almost all ERP systems are currently being migrated to a new architecture or have already been migrated to it. Some vendors (Oracle, SAP and Microsoft) use their own platforms, others - strangers. In particular, the Sage Group uses Microsoft.NET, SSA Global, Infor, Geac, Lawson and Intentia solutions use the IBM WebSphere platform, Epicor uses both Microsoft.NET and Progress Software OpenEdge. IFS Applications uses its own IFS Service-Oriented Component Architecture platform.

SAP plans to release a new version of its ERP system, based entirely on Enterprise Service Architecture and the NetWeaver platform, in 2007, Oracle plans to fully transfer all its business applications to Fusion by 2008, while at the same time it is scheduled to release a new ERP platform Lawson Landmark, Project Green by Microsoft to transfer the entire Dynamics line to the.NET platform. Thus, the application of service-oriented architecture in modern enterprise resource planning systems has long been resolved and resolved positively. The transition to a new software development and implementation model is under way, as well as the debugging of the basic functionality of new technology platforms.

SOA as "common denominator"

In parallel, these platforms are used as tools for integrating heterogeneous information systems. There are two reasons for this use: a significant number of mergers and acquisitions in the segment of large and medium-sized businesses; and the lack of funds for the introduction of monolithic ERP systems in many companies in the SMB sector.

The first reason leads to the fact that newly formed holdings, corporations, etc. often become the owners of dozens of different ERP systems used by acquired companies. Given that the lion's share of these systems works quite well, and the main thing that the holding needs is to combine information systems to quickly monitor the state of the business and obtain consolidated statements, it is much cheaper to apply an integration solution than spending tens of millions of dollars to "re-establish" a single ERP system instead of inherited ones.

The second reason is due to the growing interest of leading vendors in SMB customers. At the same time, medium and small companies are much less inclined to part with the significant amounts that are required to implement a single enterprise-wide system. If large companies can afford significant spending to "increase the transparency and manageability of the business," then medium and small are more interested in the economic feasibility of implementing holistic ERP systems and not at all against "patchwork" automation, if it saves money. In this context, the integration platform performs its primary function of providing a data interface for heterogeneous information systems. In addition, it is used in programs for the gradual transition from legacy systems to a single ERP solution of the platform supplier or its partner. An example of such a program is SAP Safe Passage. With this approach, customers can replace legacy systems with modules of a single ERP system as the first ones become obsolete or if they cease to cope with the information load. Thus, the customer's costs of implementing the enterprise system are distributed over time, which is a serious advantage for small companies, and one-time costs are incurred only for licensing and installing the integration platform.

A little theory

From a technological point of view, the transition to a service-oriented architecture of complex enterprise management systems is completely logical. Its cause, as for all examples of technological development, is contradiction. It is that an integrated system must be unified at the same time to serve as an effective mechanism for centralized business management and must not be unified to provide the necessary flexibility, adaptability and economic attractiveness of the solution. One of the solutions to this contradiction is the fragmentation of the system into parts with ensuring that their work is coordinated. Splitting ERP systems into modules was such a solution, but, due to the fact that it was carried out on a functional basis, its effectiveness was insufficient. The SOA implements a division of the system, corresponding not to functions, but to processes, so its efficiency should be significantly higher. Service-oriented architecture, as the "heiress" of object-oriented architecture, is the next step in the development of information technologies and the transition to it is dictated by objectively established need.

SOA - threat to monolithic systems

At the same time, technological development always affects the economic aspects related to the subject area. The main argument of suppliers of large ERP systems was the creation of a single information space covering all or most of the business processes of the automated enterprise. It is because of this argument that customers chose and choose large monolithic systems that allow them to get rid of "patchwork" automation and implement common standards throughout all divisions and business processes of the company. Thus, for quite a long time, the opinion was cultivated that a single solution from one supplier is better than a set of heterogeneous systems selected on the basis of the "best of breed" principle. With the widespread implementation of a service-oriented architecture, the arguments for unified solutions from one supplier lose their strength. If in the past the customer was ready to implement a single solution of the ERP system vendor in order to obtain guaranteed compatibility of all components of the information system and their integrated functionality, now and, especially, in the future, this compatibility and integration of functionality is ensured "by default" simply due to the compliance of developers with architecture standards. Moreover, if in the usual situation the customer was in a certain dependence on the vendor, which could only be eliminated at the cost (and considerable) of completely replacing the ERP system, then the service-oriented architecture allows the user to choose the provider of the service he needs, and not the system as a whole. At the same time, dependence on a particular supplier disappears with the replacement of its service with its competitive counterpart. Here, the OpenSource community is also connected, which is also actively switching to SOA.

Change is inevitable, but not catastrophic

Thus, the market strategies of ERP system manufacturers must inevitably undergo changes. At a minimum, the transition to COA should affect the number of large "monoprojects" built on the solutions of one developer.  These changes do not have to be revolutionary at all, since many companies will still prefer to work with one supplier, although in this case the developer may well replace the integrator, as is customary, for example, in the Linux world. However, the change in the economic model of vendors of ERP systems will occur gradually, since, firstly, the process of transition to SOA is quite long, and secondly, today's implementations of service-oriented architecture are still "branded" and their full standardization will also take time. Accordingly, first, competition between service developers will occur "inside" platforms from IBM, SAP, Oracle and Microsoft, and with their mutual standardization, competition will go beyond the framework. So, most likely, after the software, the economic models used by the suppliers of ERP solutions will also become service-oriented.

Sergey Sereda