EC, April 29, 2008, No M.5080

Dear Sir/Madam,

Subject: Case No COMP/M.5080 – ORACLE/BEA
Notification of 26/03/2008 pursuant to Article 4 of Council Regulation No 139/20041 

1. On 26/03/2008, the Commission received a notification of a proposed concentration pursuant to Article 4 of Council Regulation (EC) No 139/2004 by which the undertaking Oracle Corporation ("Oracle", USA) acquires within the meaning of Article 3(1)(b) of the Council Regulation control of the whole of the undertaking BEA Systems Inc. ("BEA", USA) by way of purchase of shares.

I. THE PARTIES

2. Oracle is a US publicly listed company, whose common stocks are traded on the NASDAQ. It develops, manufactures and distributes enterprise software solutions and related services, including "middleware", database and enterprise applications software.

3. BEA is also a US publicly listed company, whose common stocks are traded on the NASDAQ. BEA designs, develops, manufactures and distributes middleware software.

II. THE OPERATION

4. Oracle and BEA entered into an agreement and plan of merger for Oracle to acquire 100% of the outstanding voting securities of BEA for a total value of approximately USD 8.5 billion. As a result of the proposed transaction, BEA will be wholly owned and solely controlled by Oracle.

III. CONCENTRATION 

5. The operation therefore constitutes a concentration within the meaning of Article 3(1)(b) of the EC Merger Regulation.

IV. COMMUNITY DIMENSION

6. The undertakings concerned have a combined aggregate world-wide turnover of more than EUR 5 billion (Oracle € 13 859 million; BEA € 1 110 million)2. Each of them have a Community-wide turnover in excess of EUR 250 million (Oracle € 3 813 million; BEA € 322 million), but they do not achieve more than two-thirds of their aggregate Community- wide turnover within one and the same Member State. The notified operation therefore has  a Community dimension.

V. COMPETITIVE ASSESSMENT

5.1 Relevant product markets

7. Both parties are active in the business software market (as distinguished from consumer software)3. Business software can be broadly categorised into infrastructure software and application software. Oracle and BEA are both active in the enterprise infrastructure software segment, commonly known as "middleware".

8. Middleware refers to a wide category of software products that provide the infrastructure for applications to run on a server, be accessed from a variety of clients over a network and be able to connect a variety of information sources. This includes several products  that could constitute sub-segments such as application server software, web server software, transaction processing monitors ("TPM"), application integration, enterprise portals, event management software, enterprise service bus software ("ESB") etc.

9. Oracle and BEA have overlapping activities especially regarding the following products.

a)     Application servers4: an "app server" provides applications running on the server the functionality they need to perform a variety of tasks (communication with database, managing interaction and sessions, pooling use of system resources);

b)    Enterprise Service Bus: it is an open standard-based message bus that in principal allows applications and other software components to communicate  with  each other via a standard interface;

c)     Application Integration: category of software allowing the integration of and communication between disparate applications and information systems so that they can be used together in an efficient manner;

d)    Portals: category of software providing a framework for integrating information and processes across organizational boundaries.

10. The market investigation confirms that the middleware segment could be sub-segmented according to the end use of the product. Nearly all respondents have confirmed the existence of the different segments appointed above, while a few of them (both costumers and competitors) do not consider ESB a separate product from Application Integration, due to the fact that ESB could be considered as part of the Application Integration products and both products could belong to the same sub-segment.

11. The merging parties offer Java-language (mostly according to the J2EE standard) middleware products, just like their major competitors IBM and Sun Microsystems. Microsoft also offers middleware but based on the competing .NET platform, including Microsoft's web server and application server products. The market investigation has  clearly indicated that these products compete against each other.

12. In any event, it is not necessary for the purpose of this decision to delineate the exact definition of the product market as the transaction does not raise serious competition concerns under any alternative market definition.

5.2. Relevant geographic markets 

13. The notifying party proposes to consider the geographic market for overall middleware and the four sub-segments considered above as worldwide in scope.

14. As the market investigation has widely confirmed, customers consider offers from vendors from all parts of the world, there are no technological barriers that restrict vendors from responding to bids all over the world, and the middleware products are broadly identical across different countries, since they would not require product adaptation in order to reflect different language, legal or accounting requirements.

15. For the purpose of the present decision, the transaction will therefore be assessed under the worldwide geographic scope.

5.3 Assessment

Middleware market

16. The middleware worldwide market amounted to (approximately) between € 7.46 billion and € 9.34 billion in 2006 depending on the third-party's report considered and had a growth rate between 16% and 18% in the same year. The notifying party provided the worldwide middleware sales shares below referring to Gartner's and IDC's reports for  the years 2004-2005-2006.

17.  The parties argue that the transaction would have limited effects mainly because the market strategies and the range of products the respective parties offer are different: Oracle would be a broad solutions provider, of which middleware is only a small part, whereas BEA is (and would be perceived by customers as) a "best-of-breed"  middleware vendor, in particular with regard to application servers.

18.  This has been confirmed by the market investigation to a certain extent only. In fact, the argument may hold when the discussion is solely focused onto how their "products" are represented on their fact sheets. However, what IT vendors offer to their customers are not individual products but packages of solutions containing products, supports, consultations, etc. The fact that BEA may typically sell its offerings on a standalone basis does not negate its competitive position in the marketplace: Oracle package competes with BEA stand alone product combined (by the customer) or packaged (by a system integrator) with other products.

Data on competitive interaction in the overall middleware market provided by the parties

19. In addition, the parties provided information about the competitive interaction in the middleware market to underpin their arguments.

20. Internal reports and aggregated and partly quantitative information from internal databases illustrate the competitive interaction between competing vendors in individual software sales deals in the middleware market. The parties argue that the conclusions derived from this information also hold for the four sub-segments of the middleware market outlined above. As the parties identified the application server segment as the segment potentially most impacted by the proposed transaction, they submitted additional information specifically about the competitive interaction in the application server segment. As regards the overall market for middleware the parties provided mainly the following information:

• Oracle middleware quarterly win reports: Information about competition in relation to [150-300] Oracle middleware deals between May 2006 and July 2007, including 39 deals involving customers in the EEA. Oracle also provided quantitative summaries of these reports and an overview of selected EEA deals.
• Oracle installed base reports: For 30 EEA-based customers who bought Oracle Middleware products between May 2006 and July 2007, these reports show the scope of products for which a customer paid support and maintenance fees to Oracle.

21. Similar information for BEA’s perspective was not provided for the overall middleware segment. However, the parties submitted information from BEA’s perspective for the application server segment which will be discussed below.

22. While these submissions stem from the parties internal sources and might neither be perfectly objective nor exhaustive, they may serve as a meaningful indication of competitive interaction for middleware deals between competing vendors.

23. On the basis of the information provided for Oracle middleware deals involving customers worldwide, regardless of how many competitors Oracle faced for an individual transaction, of the […] Oracle middleware deals, […] deals ([5-20]%) identified BEA as a competitor. In the same context, IBM was involved as a competitor in […] deals ([20-35]%), Microsoft in […] deals ([0-15]%), SAP in […] deals ([5- 20]%), Business Objects5 in […] deals ([5-20]%) and Sun in […] deals ([0-15]%). This analysis does not change when assessing those deals, where Oracle faced only one competitor. BEA was identified as the only competitor in […] out of […] deals ([0- 15]%), on par with Microsoft but behind IBM ([…] deals, [10-25]%) and SAP (in […] deals, [5-20]%). Furthermore, BEA’s importance as a competitor of Oracle is not significantly higher for deals where Oracle faced two or more competing suppliers. Therefore, BEA does neither appear as Oracle’s most important competitor nor as an exceptionally frequent direct competitor of Oracle.

24. As regards competition for customers based in the EEA, regardless of how many competitors Oracle faced for an individual deal, of the […] Oracle middleware deals, […] deals ([10-25]%) identified BEA as a competitor. In the same context, IBM was involved as a competitor in […] deals ([20-35]%), Sun in […] deals ([10-25]%), Microsoft in […] deals ([0-15]%) and SAP in […] deals ([0-15]%). This analysis does not change when assessing those deals, where Oracle faced only one competitor. While BEA was identified as the only competitor in […] out of […] deals ([15-30]%), IBM  and SAP received the same score and Microsoft was identified in […] deals ([10-25]%). Furthermore, BEA’s relative importance as a competitor of Oracle is lower for deals where Oracle faced two or more competing suppliers. Therefore, BEA does neither appear as Oracle’s single most important competitor nor as exceptionally frequent direct competitor.

25. The assessment of the above information on competitive interaction indicates that BEA does neither appear as Oracle’s closest competitor nor as a frequent important competitive constraint on Oracle in the overall middleware market. Furthermore, the information presents IBM as the most important competitor of Oracle, and shows that Microsoft, SAP and SUN regularly compete as vendors of proprietary middleware.

Results of the market investigation

26. The market investigation largely confirms the conclusions drawn from the data submitted by the parties. In the overall middleware market, respondents do not consider Oracle and BEA to be close competitors. Indeed, they see that if Oracle is seen as a competitor of BEA and vice versa, they are not the closest one to each other.

27. The competitor of each of the parties that is perceived as the closest of each one in the overall middleware market is IBM. It could furthermore be noticed that when IBM is  not quoted, competitors or customers consider other competitors (such as SAP, Sun, Microsoft or Tibco for instance) closer to one of the parties to the transaction than the other.

28. The market investigation also indicates that Microsoft in all four product markets acts as a competitive constraint, which a majority of both competitors and customers reported.

29. Open source software providers (e.g., Apache, Red Hat/JBoss, Geronimo etc.) are by a majority of competitors and customers seen as an alternative to the parties' products. While the market investigation showed that some customers refrains from using open source middleware for mission critical components, it indicated that open source solutions have matured to a point where they can now be considered viable alternatives to traditionally licensed products for certain tasks.

30. Some players still consider that these open source alternatives offer similar functionality to  a certain extent only, as there are limitations to their functionality and support/certifications that traditionally licensed products address. But it must be admitted that products from open source software providers (such as JBoss) could be within mission critical environments dominated by commercial vendors so far, and concern both SME as well as large companies.

31. This competitive constraint is further strengthened as more and more companies are choosing to leverage open source solutions as part of their IT strategies in order to reduce license costs and provide their organizations with more control over software development and direction; they are evaluating commercially license middleware against these open source solutions. Finally it is usually seen a mixture of open source and closed source elements in the IT environments of many organizations and not just one solution.

32.  As regards the impact of the merger on the overall middleware market, it results from the above that customers will continue to have a sufficient choice among alternative vendors of middleware software with IBM, Sun, Microsoft, SAP and Open Source vendors in full participation. One third party raised the concern that Oracle's products are bundled and there would be no space for competitors to compete. However, already now, other competitors such as IBM (which also provides middleware on a standalone basis) and SAP are able to provide middleware together with different product. Therefore, the middleware market would not lead to see the merged Oracle/BEA entity to gain market power and hence no competition concern would arise from non-coordinated effects stemming from the proposed transaction.

33. Furthermore, in view of the specific characteristics of software markets, the merger would not be conducive to coordinated effects either. Should the new entity and IBM behave in such a way, numerous competitors as listed above (SAP, Sun, Microsoft and open source providers) would have the means to exert a competitive constraint. In addition, middleware solutions are differentiated products, and their pricing is non- transparent, also because of the number of vendors and their practice to apply discounts on public prices which are agreed typically on bilateral negotiations. As a consequence, price comparison between similar products is difficult. Therefore, the proposed transaction is unlikely to give rise to any coordinated effects, in particular between IBM and the merged entity.

Application servers

34. The notifying party provided worldwide market shares for this segment based on  Gartner and IDC reports.

35. The significant discrepancy between Gartner's and IDC's conclusion, with respect to Oracle's position, reflects, in the notifying party's view, the analysts' application of different methodologies and revenue allocation mechanisms to the same vendor product revenues. In the notifying party's view, on the one hand, Gartner underestimates Oracle's position as it does not take into account application server revenues obtained as part of a broader product bundle. On the other hand, IDC overestimates Oracle's position by counting revenues that are not purely application server-related. Also BEA's position would be overestimated as BEA's license revenues have declined in the years 2004- 20067.

Data on competitive interaction in the application server segment provided by the parties

36. As discussed above, in addition to the above market shares and arguments, the parties have provided internal reports and aggregated and partly quantitative information from internal databases to illustrate the competitive interaction in the application server segment.

37. As the parties identified the application server segment as the segment potentially most impacted by the proposed transaction, they mainly submitted, in addition to the information discussed so far, the following information concerning the application  server segment:

• Oracle Sales Online: Oracle provided an extract from its database covering all sales opportunities involving application servers in Europe, Middle East and Africa between September 2005 and November 2007. The database is an aggregated record of [thousands of] sales opportunities and includes  information on known competitors.
• BEA runner-up survey: BEA submitted data from a survey about the competitors that were identified as runners-up on large BEA deals8 that  included an application server. This data includes information on competitors in [150-300] deals surveyed between 2006 and 2008.
• Oracle and BEA top 50 deals: Reports on the basis of individual deals showing the competitors identified for Oracle and EEA deals, respectively, involving an application server in 2006 and 2007.
• BEA uncontested deals: BEA provided summary data on the total number of its application server customers and deals where it faced no competition between 2006 and 2008.

38. The assessment of the information from Oracle’s Sales Online database leads to the following main findings: As for the overall middleware market, IBM is identified as the most important competitor in the application server segment. According to the number  of mentions received in the database, IBM is the leader with [10-25]% of mentions followed by Microsoft with [5-20]%. BEA appears in third place with [0-15]% followed by SAP ([0-15]%) and SUN ([0-15]%).

39. Moreover the data point to the existence of competition from open source vendors, the existence of smaller vendors, as approximately [10-25]% of the deals named  competitors other than BEA, IBM, Microsoft, SAP or Sun, and a substantial share of uncontested deals ([45-60]% of application server deals for which competitor information was indicated).

40. The assessment of the information from BEA’s runner-up survey shows that while IBM was mentioned as a competitor in […] deals out of the total number of […] application server deals ([25-40]%), Oracle was mentioned in […] transactions ([10-25]%) and open source in […] deals ([0-15]%). Moreover, the data also points to competition from a number of smaller application server vendors as well as to a substantial share of uncontested deals ([20-35]%).

41. As regards the assessment of Oracle’s top 50 deals involving application servers, the main findings are that IBM, Microsoft and SAP regularly compete with Oracle. While BEA is also mentioned in a number of those deals, BEA appears neither as a very frequent nor as an important direct competitor of Oracle in those deals.

42. The assessment of BEA’s top 50 deals involving application servers also underlines the leading position of IBM with […] mentions in 2007 ([35-50]%) and […] in 2008 ([25- 40]%). While Oracle is an important competitor with […] mentions in 2007 ([5-20]%) and […] mentions in 2008 ([15-30]%) it remains second to IBM. Furthermore, smaller vendors appear as significant competitors with a share of […] in 2007 ([0-15]%) and […] deals in 2008 ([5-20]%).

43.  The assessment of BEA’s uncontested sales shows that the ratio of deals without competition  in  BEA’s  total  application  server  deals  has  in  general  increased. This suggests that for those deals BEA mainly sells to its existing customer base where there is no competitive interaction.

44.  In sum, the information assessed above appears to support the various contentions of the parties. Oracle and BEA do not appear to be each other's closest or most important  direct competitors in the application server segment. IBM, the market leader, as well as Microsoft, SAP and SUN would remain significant competitors post-transaction. Open source vendors would continue to exert competitive pressure on the merged entity at  least for some deals. Furthermore, a significant proportion of application server deals were uncontested in the past and BEA’s importance for competitive transactions seems to be declining. Consequently, the transaction data confirm the parties’ statement that  the transaction would not lead to competition concerns based on non-coordinated effects in the application server segment.

Results of the market investigation

45.  The market investigation has confirmed these statements. It appears that neither Oracle nor BEA should be considered as the closest competitor of each others in the application server is segment. The same findings as for the overall middleware market are valid here again: IBM appears to be the closest competitor of each of the parties. And when one of the parties is named as a competitor to the other, it is worth noting that other competitors (such as the open source provider JBoss) are also indicated as suitable competitors.

46.  It seems that within the application server segment the offering of each party would be perceived differently in the customers' view: BEA produces best-of-breed application servers on a stand-alone basis or as a central element of the deal with a particular focus on selling additional licences9 to its installed base. Oracle's application server would be  a comparatively new product within Oracle's extensive enterprise software portfolio, almost always sold in the context of a wider package of infrastructure technology or application software.

47. Furthermore, Microsoft acts as a competitive constraint. Microsoft's .NET middleware, including in particular Microsoft's web server and application server products, is  bundled with the Windows Server operating system and is not licensed or priced separately. Additionally, the other Microsoft middleware products are marketed and  sold under the Windows Server product family and typically purchased along with other Windows Server products. The Windows Server operating system that bundles .NET middleware has around 70% share of the overall server market. In other words, on average, around 70% of all middleware opportunities involve the customer having the option of using the Microsoft application server functionality already acquired as part of their Windows Server operating system. In this way Microsoft acts as a significant constraint on middleware vendors.

48.  Competition stemming from Microsoft is utmost true at the stage where a customer is planning the installation of his IT environment, and then having to select standard-based (such as Oracle) or a proprietary (such as Microsoft) system. It remains true as a majority of customers indicated that they use Microsoft's .NET and Java technologies side by side for similar tasks (which would allow switching from one system  to  another). A large majority of competitors also confirmed this for their customers. The reasons for running both technologies in parallel seem to vary widely as does the customers' assessment of whether or not such parallelism is desirable or a mere necessity. A minority of customers indicated that they run both technologies in parallel only for historical reasons, such as the merger of two formerly independent companies with separate IT environments, but that they would prefer not to. A number of customers indicate, however, that each technology has its specific advantages.

49.  As regards SAP's NetWeaver, only a slight majority of customers reported that they use both NewWeaver and Java technologies side by side for similar tasks, even though a large majority of competitors expected their customers to do that.

50. As indicated for the overall middleware market, open source software providers also exercise a certain competitive constraint in the application server segment.

51. It results from the above that customers would continue to have a sufficient choice among multiple alternative vendors of application servers. The new Oracle/BEA entity would therefore not give raise to serious competition concern.

Portals 

52. In this segment, the market leader post-merger would be IBM, with a market share of approximately [25-35]% in 2006 (both for Gartner and IDC), followed by the merged entity with a market share of around [25-35]% (split in [15-25]% to BEA and [5-15]%   to Oracle both for Gartner and IDC). Microsoft, whose market share in 2006 was for Gartner [15-25]% and for IDC [0-10]%, would constitute the third strongest competitor. Once more, the discrepancies between the figures provided by the analysts reflect, in the notifying party's view, the wide underreporting issues connected to Microsoft's middleware "free-of-charge" offer.

53. The findings from the assessment of the transaction for the overall middleware market and for the application server market also hold for the portals segment: Oracle and BEA should not be considered as close competitors. IBM, which would benefit from a wide range of complex deployment patterns thanks to its large customer basis, appears to be the closest competitor of each of the parties. And when one of the parties is named as a competitor to the other, it is worth noting that other competitors are also indicated as comparable competitors. For instance, the market investigation has confirmed that Microsoft, which would have a competitive advantage drawn from its presence in the Operating System and server markets, and in the markets of tied products  such  as Office, has the ability to exert a significant competitive constraint. Consequently, the proposed transaction would not give rise to any competition concerns in the portals segment.

ESB

54. In the Enterprise Service Bus market, according to Gartner in 2006 IBM is the market leader with a market share of [35-45]%, followed by Tibco with a [10-20]% market share, Oracle with [10-20]% and BEA with [5-15]%. Together, the merging parties would hold a market share of [15-25]%.

55. The findings from the assessment of the overall middleware market also hold for the ESB segment.

56. Furthermore, according to the market investigation, the ESB segment is a constantly growing and dynamic segment, in which standards have not yet been established. For  the latter reason, this category would encompass products which are comparable only in very large terms. In this respect, BEA's and Oracle's ESB offerings substantially differ  in their technical characteristics and in the way they are commercialised (BEA' standalone product "Aqualogic Service Bus" vs. Oracle's larger suite bundles, including Integration and Service Bus and App Servers).

Application Integration

57. In this segment, according to Gartner, the parties' combined market share would be [25- 35]%, with Oracle holding a [20-30]% (market leader) and BEA adding [0-10]%10. According to the notifying party, Gartner includes in this particular category an "arbitrary" selection of products, which are only comparable in broad terms. In Gartner's own words, products that are included in this category "may exhibit integration suite or application platform suite functionality or both".

58. The findings from the assessment of the overall middleware market also hold for the application integration segment. Beside the limited overlap caused by the transaction, it can also be noted that major vendors of software and various specialized players offer suite products that include application integration in addition and that several vendors also offer open source integration suites. Finally, because of the specific characteristics of software markets mentioned above mainly relating to the lack of transparency, the merger would not be conducive to coordinated effects.

59. As regards the impact of the merger on the overall middleware market as well as on   each segments presented above, customers will continue to have a sufficient choice among multiple alternative vendors of middleware software and among alternative vendors of, respectively, each of the four product categories.

VI. CONCLUSION

60.  For the above reasons, the Commission has decided not to oppose the notified operation and to declare it compatible with the common market and with the EEA Agreement.  This decision is adopted in application of Article 6(1)(b) of Council Regulation (EC) No 139/2004.

¹       OJ L 24, 29.1.2004 p. 1.

²  Turnover calculated in accordance with Article 5(1) of the Merger Regulation and the Commission Notice on the calculation of turnover (OJ C66, 2.3.1998, p25).

³  See Commission Decision in Case COMP/M.3216 – Oracle/Peoplesoft of 26/10/2004.

⁴   Oracle offers three principal products that incorporate application server functionality: Oracle internet Application Server Java Edition, Standard Edition and Enterprise Edition. BEA's application server product is Weblogic. BEA derives [65-80]% of its revenues from the sale of Weblogic and of Tuxedo, its TPM product.

⁵   Acquired by SAP.

⁶   Gartner’s market share report does not take into account any Microsoft application server presence.

⁷   Annex 16 to the Form CO.

⁸   BEA only surveyed deals where revenue exceeded USD 250,000 in 2008, or USD 500,000 in 2006 and  2007.

⁹    According to the notifying party, as indicated above, a significant portion of BEA's app server revenue derives from sales that are not made in competition with any other app server vendor, as they are derived from additional licenses sold to its customer base.

¹⁰   According to the notifying party the Oracle and BEA products that Gartner uses in this analysis do very different kinds of Application Integration. Specifically: Gartner includes four primary BEA products in  this analysis – AquaLogic Data Services; AquaLogic ESB, AquaLogic BPM, BEA Mainframe  Integration, and WebLogic Integrator. Gartner includes the Oracle SOA Suite which includes several different products in its market share analysis of Oracle’s market share. These product suites are fundamentally different in three important ways: (i) Oracle’s product family includes a number of components that BEA does not provide; (ii) BEA’s product family includes a number of components that Oracle does not provide and (iii) in the two areas where there could be perceived overlaps – ESB and business process management ("BPM") – there are significant differences between the corresponding Oracle and BEA products   which do not make them substitutes from the market sense.