September 20, 2012

Best of Breed or Vendor Lock-in, isn't there something in the middle? Yes. Ecosystems.

The CIO Challenge

If you have a budget and you need to decide on your company's IT strategy for the next 3 to 5 years, you might have been drawn to this very same question: "Should I choose Best of Breed systems, or should I align with some vendor's strategy taking the «vendor lock-in» risks that come with it?".

If you have ever asked this to yourself, this article is for you. If you're not a decision maker but influence long term decisions in the technology area, probably acting as a CTO or Chief Architect, then this article is also for you.

"Your opinion is biased: you work for Oracle": Let's bury the agendas for a while.

I am not an analyst, I work for a vendor: Oracle. So is expected that my words are filled with biased intentions, second intentions and a secret agenda. Let's step back a bit. Let's imagine that I don't work for Oracle. Imagine that I'm an analyst working for Gartner, McKinsey, Forrester or any other market analyst firm. Would you expect that my words be less biased? What would be my agenda then? Let me tell you. My agenda could be to beat the crap out of every vendor, in order to gain the CIO's trust, so he could buy my studies and seat tickets at my conferences. So analyst or not we all have an agenda.

Does this mean that you, Mr. CIO, are completely alone in this quest because like the richest man on earth, you can't distinguish people's behavior around you floating between real interest in your problems or wanting to rip you off? Of course you're not alone, because before being a CIO you probably worked for, specialized in, or consulted for some of these vendors. So you too have an agenda apart from stretching that IT buck and looking good in the executive board picture.

What I suggest for the next few paragraphs is that we all put our agendas aside and concentrate in this fundamental problem of choosing one technology over the other. People seem to have an opinion about every technology even the ones they never experienced. How? Because they read articles from experts or the market analysts, so they don't need to have real experience to choose. They read studies, and after a thourough analysis the strategy seems to fall for one of the sides: "Best of Breed" approach, or the so called "Vendor Lock-in". But I'll show you my own analysis, biased or not, that points to something in the middle called: "Ecosystem".

The origin of "Best of Breed"

The "Best of Breed" approach came when big computer systems started to be offloaded from the companies like IBM or DEC to the Unix based offerings of those very same companies, or onto other platform from players like HP or Sun Microsystems. All of the sudden you were not buying whole systems anymore from your vendor of choice, but going into an approach where the storage systems came from one vendor, the operating system from another, the application development technology from another, the database from another and so on, and so forth. This was the beggining of the era of Open Systems. The era of "experts". We all know that "experts" are people who know a lot from a very specific technology or area. Companies did not ship loads of black boxes anymore (with free software and free services), but instead new players came that were manufactoring systems with disks from one vendor, chips from another, operating system from another, database from another, and networking from another. Now the buyers had to figure out who was shipping the most attractive system for the least amount of money. Buyers had to start being "spec experts". So buyers had to figure out which company was making the system that assembled the best constituent parts, the so called Best of Breed solutions. This approach extended to the enterprise software as well when business software started to be sold as packages in the mid 1980's. Enterprise Resource Planning (ERPs) being one of the first market trends fighting this Best of Breed approach. Even though, these software packages had to be tested against the plethora of systems combinations that existed back then, so the buyers life would be much easier. The software vendors had to close down partnerships with the main systems vendors, so before a new system was released to the market, it would be already "certified" with the main software vendors. But this was as far as the "Best of Breed" approach could go in software: make sure all the pieces you buy, would fit together and that no explosion would happen when you start connecting them. The real catch is that none of the software vendors nor the system vendors could pre-test all the possible combinations, and eventually problems arose that made way for loads of manual work in integration. This was the beggining of the era of System Integrators (SIs), with their armies of experts and project managers, later called "consultants".

Why Best of Breed failed

This approach started showing signs of failure in the 90's, but some IT departments  would still insist on it because the alternative was going back to black box systems, the ultimate vendor lock-in, and the huge costs associated with that. On the other side of the counter vendors knew something that their business customers didn't: the market couldn't afford to have so many IT vendors. Those were the days where the Oracle Database was developed for more than 90 different platforms! This variety of platforms had to come to an end. So the IT Vendor consolidation era began. All, absolutely all of them had one same enemy: IBM. All of them wanted to be the alternative platform to IBM's ruling systems. A platform that could serve the customers with all the pieces, but at the same time would still give them the illusion of the Best of Breed ages. IBM was on that race as well, but its size was the main blocking force. Also "IBM was unable to pick a market in which they wanted to compete. Customers were becoming more and more fragmented, and IBM wanted to provide for everyone. However, in trying to accommodate everyone. IBM accommodated no one." says Prof. R. Preston McAfee about those times.

The first stunning movement was when outsider COMPAQ acquired Digital Equipment Corporation (DEC) once the second largest company in the IT industry. IBM was also moving forward in both the software arena (buying Tivoli) or hardware arena (with the acquisition of Sequent), while HP was still very much focused on hardware acquisitions an example being Convex. Right before COMPAQ made its move on DEC, Oracle Corporation had stepped in doing a very strategic and important acquisition: DEC's Rdb database engine. It's not until 2001 after Oracle went into a massive internal restructuring, that Larry Ellison started to put in practice a software consolidation strategy, starting with a large one: Peoplesoft for $10.3 billion. Back in the 90's, Microsoft's acquisitions strategy could be sumed in: before and after the internet. Fox Software being the most important one in the early 90's (1992) and the first global scale free email service, Hotmail for $500 million in 1997. They were creating their own technology, developing in house the next generation development platform that could compete with Java: dot net (.net). Larry Ellison once referred to it as "not yet" since Microsoft was struggling with leaving the Visual Basic environments. Bottom-line: the 90's were a fantastic age for System Integrators, because although the mergers and acquisitions movements started in that decade, it really didn't meant nothing practical for the IT departments of companies worldwide.

Double headed Age

Just when the 2KY madness ends and loads of money were thrown at making sure that mankind wouldn't be sucked in a black hole dropping us all in 1900, HP makes a move into  becoming the new IBM. HP merges with COMPAQ creating a monster company. With this move, the message was clear: another vendor was lining up for the return of the Vendor Lock-In era! If people were fed up with IBM they now could reach out to HP who had it all, including a strategic aliance with the main software vendors like Oracle, Computer Associates (CA) and Red Hat. In some cases HP didn't even waited for the big IT departments to reach out, they acquired some of those IT departments offloading companies from that burden of maintining a costly IT department: Procter and Gamble being one of those acquisitions for the total of $3 billion. A new kid was on the block and it seemed to have everything to upset the mighty IBM.

Big Blue's response was showing the market that it had offering outside their own closed locked-in environments. IBM claims it invested more than $1 billion in 2001 promoting the use of Linux and the Informix acquisition in 2001 is another example of that strategy. HP was partnering up with Oracle, that had surpassed Informix long time ago, so the acquisition from IBM was not just wrong but also late. Another example of IBM's mistakes that led way to HP growth, was betting in Enterprise Integration trying to solve a problem that was created by the Best of Breed approach that so many customers embraced when running away from Big Blue in the past.

IBM was acting as a software and services company, leaving a blank space in the market for HP in the hardware market. By the time IBM had passed the 50% mark in revenue coming from services, HP made another move, this time were the money was: consulting services. HP bought systems integrator EDS for $13.9 billion. Another example was the massive effort of other companies like Fujistu and Hitachi to be major players in the services arena. A proof that the Best of Breed approach failure meant more money then ever.

The last table spin: Oracle acquires Sun

In 2009 the acquisition spree that IBM and Oracle went on trying to fight in every layer of Information Management software, had come to a point where the market was pretty much defined already:

1. IBM and Oracle: leaders in database, middleware, service integration, business intelligence, and enterprise content management. Microsoft bitting in some markets for the third place, and even able to upset the blue and red companies, specially in the business intelligence market.

2. Oracle and SAP: leaders in enterprise applications

3. IBM, HP, EMC and Cisco: leaders in enterprise hardware systems

So in the acquisition of Sun Microsystems by Oracle in 2009, only concluded in January 2010, resided the last big market shift that was supposed to bring another big hardware player into the market, where Oracle was an outsider and Sun had lost their former glitter. Some very short sighted analysts were saying that Oracle was in it only for the Java programming language, but soon the market realized that Oracle was another runner up for the IBM's spot, a bit like HP after the COMPAQ's acquisition. But these were different times in 2009: IBM had sold it's PCs division to Lenovo, HP was still merging with EDS (in a self discovery exercise as well), and Oracle had to spin the tables and bring something in between the best of breed approach and the ultimate vendor lock-in which came to be known as Engineered Systems. I'm still not wearing my hat of Oracle employee here, but seems to me that what Oracle was targeting a new and fresh approach: the creation of that mighty Ecosystem, which is not a vendor lock-in nor a heterogeneous Best of Breed approach. All other big IT vendors kept on insisting on innovation as well, but of their current products. Oracle was creating new and revolutionary products. In trying to stop Oracle's increasing success with the first Engineered Systems (Exadata), IBM makes a move into buying Nettezza and EMC another one into buying Greenplum, but it's SAP that makes the genius one by acquiring Sybase the missing link in SAP's software portfolio: a strong database engine. No one changed their business model more than Oracle by coming up with integrated solutions where their software runs optimized, creating innovations that you could only take advantage if bought in bundle: hardware and software engineered together.

It's the beginning of the era "Enterprise Ecosystem" or Cloud if you will. In cloud it's the apps that matter, but if you want them to run seemingly , there has to exist a platform under the covers that complements that Ecosystem. This is where Oracle seems to have all the elements to close the circle. The comparison between Oracle and Apple's approach in the hardware and software integration for consumer electronics, is used ad nauseam to illustrate what Oracle is trying to create in the enterprise computing arena. By 2012 Oracle has 7 (seven) Engineered Systems in the market addressing all layers of enterprise computing: storage, database, middleware, applications, big data and business intelligence.

Other moves at integrated systems are already in place by IBM, HP, SAP, EMC and Cisco, trying to create that Ecosystem. Initiatives like VBlock, PureSystems, and HP/Violin will make the next years very interesting to watch. So, Mr. CIO the name of the game is Ecosystem and if you are still living in the Best of Breed days you will have a hard time adapting. (Now putting Oracle's cap on!). But make no mistake: Oracle is way more advanced than its competitors in this race. (Cap off).

Where does Big Data fits?

If you read until here and are wandering "Where does Big Data fits in?". Big Data technologies are the "revanche" of the open source community headed by Apache that saw it's incubators in such companies like Yahoo!, Facebook and LinkedIn. Now this has came to bite the "traditional" IT vendors in the @ss. Soon another potential future giant will emerge from the Big Data vendors, and the whole merges and acquisitions bubbling will either smothered it or clone it. It's all open source in the end, so buying companies without proprietary software doesn't seem to make the day among of the big IT vendors out there. So when the Mr. CIO is faced with Best of Breed, Vendor Lock-In or Ecosystem approach he might think that Big Data technologies are a forth way. I'm sorry, but it's not, it is complementary to all of them except the vendor lock-in because no one can span to all levels of Big Data technology needed to tackle the huge challenges that information overload and exploration will pose in the next few years. Challenges that include security, BYOD, Mobile, desktop, cloud apps brokerage, etc.

The name of the game is: build an Ecosystem and don't worry about "best of breed" or vendor lock-in myths. An ecosystem is so much more than something that works for your company, that aligns with your business initiatives, and that gives you the best cost/performance ratio with flexibility. An ecosystem is about IT becoming invisible, or even better moving away from being a cost center to be seen as a profit center.

The views expressed on this blog are those of the author and do not necessarily reflect the views of Oracle.


  1. "The name of the game is: build an Ecosystem and don’t worry about “best of breed” or vendor lock-in myths. An ecosystem is so much more than something that works for your company, that aligns with your business initiatives, and that gives you the best cost/performance ratio with flexibility. An ecosystem is about IT becoming invisible, or even better moving away from being a cost center to be seen as a profit center."

    I dig it. Furthermore, nowadays it is very likely that the ecosystem you choose is comprised of best of breed technologies, all wrapped up in a pretty VBLOCK bow.

  2. Reblogged this on Eyetdk's Blog and commented:
    Interesting thoughts, although a bit Oracle biased, on the “Enterprise Ecosystem” :-)

  3. Thanks for the reblog Kristian! You are absolutely spot on. When I read that paragraph again I saw that the bridging between the ecosystem and the cloud computing is a bit of a stretch, but I didn't wanted to go onto cloud computing concepts, which actually are less related to technology and more related to business.

  4. You are right Dan: any ecosystem will not only used best of breed technologies, but also use bigger building blocks. Not sure if virtual or not, but most surely bigger :-)