Sybase Database Value to SAP – Long Term and Short

It’s not what you think – the hidden jewel for the near term may just be SQL Anywhere. Read on. Disclosure: I worked at Sybase in the last millennium, when it hit the wall at $1B the first time and bounced. Over the next few years, Oracle dramatically outdistanced itself, in large part, as it turned out, because of the massive opportunity presented by SAP. Thousands of huge installs atop the Oracle DBMS, and not one with Sybase. Why? Because of a technology disagreement. SAP wanted row-level locking. Sybase’s answer: “Let us tell you why you’re wrong to want it.” Leaving aside the lesson to be learned from that one, let’s talk about how much the newly acquired Sybase database portfolio does for SAP. I’m leaving the best for last, because all the chatter has been about ASE and IQ, but read to the end.

First: will Adaptive Server Enterprise (ASE) have an impact on SAP revenue anytime soon? Can SAP use ASE to impact Oracle? SAP is struggling to keep pace as the chart here (from the Economist post) shows, and any front in a broad battle is useful. So how much direct impact will ASE have?

  1. Not on the ERP side. Porting to ASE will take SAP time. It can be done;  the fact that it runs on SQL Server, built on code Sybase provided to Microsoft while I was there, makes it clear that it can be done. [edit 5/14 removed reference to SQL Server’s installed base] And once it’s ready to go, how soon will customers port from Oracle to Sybase? That will be years, literally. If they decided to start the project the next day it became available, the port would still take a long time, cost a lot of money and resources, and disrupt many things. When they get their Oracle renewal numbers, they might then assess the costs – and still decide that it will be too expensive. This applies to Microsoft too, but the license/maintenance differential will be even less of a motivator there.
  2. What about ByDesign? Same answer. Business ByDesign, SAP’s in-memory play whose new version is to be unveiled next week, is a different code line and a bigger volume play than
    Stock price chart
    Price gap widens

    traditional ERP going forward, and Paul Hamerman at Forrester thinks this might be the way SAP might go, though he doesn’t mention it in his blog post specifically. Only time will tell, but SAP needs to stem the erosion and grow its mid-market base, and this may be the play. If ASE’s in-memory technology, which is relatively new, can make an impact here, or, for that matter above – SAP’s own much-discussed in-memory play is mostly possibilities, not product, today – that might have an impact. But again: soon? No.

  3. How about BW? IQ maybe, but ASE, No. [edited 5/14, removing factual error] SAP is close to delivering on its “BW on Teradata” project, soon to come to market. That project has been underway over the past two years; there has been a great deal of investment in it, and pilot customers are already in the wild. BW also runs on Oracle, DB2, and SQL Server, with its default being SAP’s own MaxDB, the result of work with Software AG some years ago. Where will Sybase’s leading analytic DBMS, Sybase IQ, fit? That’s not yet clear, but it has enormous upside in SAP accounts if it’s sold aggressively. It’s done very well despite a relatively quiet marketing effort over the years, which I have often talked about, most recently here. I believe that play will happen fast, and IQ will be a supported choice fairly soon; its relationship to MaxDB is to be determined. ASE? Not so much.
  4. Will ASE make money anyway? Yes. It has continued to grow for the last several years, adding new name accounts – and not just in China, although it’s done quite well there. SAP pointed out on their public call yesterday, Sybase will make money for SAP even before synergies click in, and ASE is one reason.

Second: what about the Sybase IQ opportunity? Here, as I said above, the opportunity is far more obvious, and I suspect we’ll see action soon. IQ is the most widely installed columnar engine for analytics in the marketplace, the most mature, and the one to beat. In my discussion with Vishal Sikka, I asked about the opportunity to use Sybase’s powerful heterogeneous data movement capability with its industry leading Replication Server product, propagating data via its new direct connection to Sybase IQ. And putting Business Objects in front of that. He lit up. “You get it,” he said. Pricing and packaging? Integrated bundles? Of course. The possibilities are rich and complex – and they will take a while. But that is a big play.

Third – how about mobility? Yes. This is the big one. SQL Anywhere, in play in 81 of the Fortune 100, has been the leading mobile database for years, with 10 million deployed seats. It’s already a big OEM play, and with the new release 12 about to ship (beta download here), it adds functionality that SAP will use to good advantage, as will the 1500 partners that embed it in their offerings today (companies like Cisco, Symantec and Intuit, for example). SQL Anywhere already claims some 20,000 developers who will continue to build on new features, developing, say, data-driven spatial applications for smartphones, synchronizing on-device spatial data with spatial databases, application servers and enterprise systems – huge opportunity within the SAP base. SQL Anywhere has iPhone support in beta now; one can easily see the fit there – not to mention the iPad opportunity. It’s already in the right spaces: mobile commerce, utilities, telecommunications, manufacturing and government – all core SAP markets. In terms of immediate available impact for SAP’s applications, we’re likely to see this happen earlier than anything else. And it could be massive.

Some great posts: Dave Kellogg, Curt Monash, Dana Gardner, Dennis Howlett, show the skepticism and concerns we all have about execution, and raise points about important dimensions. They’re all well worth your time. [ edit 5/14 – add Jim Kobielus =- a very solid piece]

Disclosure: Sybase and SAP are clients of IT Market Strategy.

Published by Merv Adrian

Independent information technology market analyst and consultant, 40 years of industry experience, covering software in and around the data management space.

10 thoughts on “Sybase Database Value to SAP – Long Term and Short

  1. We both worked at Sybase at around the same time, I support your view that Sybase IQ could be big, but I don’t think they will pull it off. Here are my thoughts on the topic.

    There is a lot of buzz surrounding SAP’s just-announced acquisition of Sybase. I’ve been asked a lot of questions about it — what it means for the column-based database market in general and SAND CDBMS in specific — and thought I would share my take.

    SAP is a smart company. The acquisition of Business Objects was a good decision and well executed, and this latest move has the making of another well-timed purchase.

    However, you can’t change a company’s core business overnight. It took Sybase over a decade to move from being a leading transactional relational database vendor to being a leading mobile software company. SAP’s business is rooted in enterprise software applications, but they know the future of enterprise software is in mobile. Many other companies are seeing this trend and are also scrambling to move their applications onto mobile platforms.

    The ongoing disruption of modern computing is moving toward ubiquitously connected, very personal, mobile devices. Massive, cloud-based data stores feeding fully capable mobile devices with easy-to-use software is the future. Salesforce.com have done a great job of connecting the Enterprise Cloud with Enterprise Mobile. But Salesforce is also a relatively simple application with a homogeneous user profile, a recipe that has powered phenomenal growth. If only SAP were this simple.

    SAP’s applications are highly complex and their users highly diverse. SAP’s core strength — managing tremendous complexity — has become their biggest challenge. Moving SAP into the cloud and connecting it to mobile applications is a major undertaking, but is the future of SAP. The acquisition of Sybase and their mobile suite of applications is a smart next step on this journey.

    The question SAND has now been asked is this: With the drive toward mobile and Sybase’s stable of mobile products, what of Sybase IQ?

    Sybase’s acquisition of IQ never really delivered any benefits. Sybase’s core business at the time was transactional, with sales and marketing arguments all about why transactional worked for everything.

    They simply didn’t get column-based databases.

    Under the leadership of John Chen, Sybase became a great mobile applications company, but IQ never made it either fully into the cloud or on mobile devices –- it was left hanging.

    SAND is laser-focused on the belief transactional relational structures are simply the wrong data structure for analytic applications.

    We’ve been proving this for the last 15 years. Driven by our customers and the genius of our visionaries, Ted Glaser, Arthur Ritchie, Richard Grondin, and Mike McCool we have taken the concept of column-based databases and built the most scalable, flexible, robust and easy-to-use column-based database in the market.

    SAND is massively parallel and mobile deployable. SAND scales in the multi-hundred terabytes with 10,000+ users on one application.

    (Yes, I said “on one application” not aggregated across instances like the numbers typically used by our competitors.)

    SAND delivers scalability on demand, in the cloud, and with native connectivity that drives deep data integration. We’ve been doing this for more than a decade while others were sleeping or just getting their companies off the ground.

    Perhaps SAP will re-purpose IQ or combine it with Sybase’s comparatively under-funded transactional relational product (remember that?) as they square off directly against Oracle. I believe SAP have a lot to do simply unpicking the myriad Sybase products and integrating them into the SAP product line of enterprise applications.

    Already I have been fending off questions about what acquisitions this move will spark in the columnar space. I think none for now. SAP did not buy Sybase for IQ. The columnar market is growing like a weed, but isn’t on the radar of Oracle, IBM or other potential acquirers. Transactional relational vendors are throwing every argument, code line, and technology acquisition to create Frankenstein databases that the smart villagers are chasing out of town by torch light.

    SAND customers know if you want scalable analytic data in the cloud for large data sets and access on mobile devices you need the right architecture for the task.

    You don’t take a Hummer on a race track, and don’t use a transactional relational database when you need a Ferrari.

    1. Mike, although self-serving, this is a thoughtful comment, and I very much appreciate your taking the time to join the discussion. I agree that IQ could – should – have been much more than it is by now. I was part of the team that first brought it to market and then, as now, it was not top of mind, or aggressively funded and promoted. As I’ve said here before, its success – and with thousands of installs and annual new customer numbers that are greater than most of its new competitors’ totals, it has to be considered successful – has been constrained by continuing timidity in the marketplace. Analysts know (it’s not a secret), but the rest of the world doesn’t see it much.

      You’re not alone in observing that SAP will have much to do in bringing the Sybase portfolio into its own, integrating it and leveraging the opportunities. And of course, it’s far too early to tell exactly what they will do, and when, and how well they will succeed – it’s all speculative. For now, at least, I have to agree with their assertion that Sybase will make them money immediately. The accretive revenue is not trivial, and it’s a benefit before synergies kick in. And IQ will play a part in that.

      One could make the same comments about marketing when discussing SAND, as you and I have in recent conversation. I’m looking forward to the impact you expect to make bringing it back out into the light again in the months ahead. Don’t be a stranger.

  2. When I read of SAP joining forces with Sybase, my first impression was that of an ERP company joining forces with a database company. SAP has the many partnerships across the database community already, so the news was less interesting until reading the points raised by Merv and Mike.

    Merv and Mike thanks for offering insight into the dimensions of mobile applications and industrial strength cloud scalability.

    From my perspective I will add a few thoughts also adding strength to the Sybase SAP partnership.

    My observation is that some database architects are proposing transactional data, such as is found in ERP databases, be the same data format as the analytical databases or data warehouses. This thinking extends out to having a single database, a single data model and a single copy of the database serving transactional and analytical purposes. For a thoughtful paper on the topic see Barry Devlin’s work. For now, let’s call this “single copy theory”.

    Where does this play out for IT buyers and SAP?

    Should “single copy theory” emerge as an accepted principal, the adopting IT buyer significantly reduces the complexity of their infrastructure, hardware, software, human resources, while making transactional data immediately available for analytical processing. The value to operational dashboards and real-time analytics is immense.

    Before the villagers chase me out of town with torches, some of these systems are in place and being used. I could feel the heat of those torches. Some IT architects never knew not to put transactional and analytical systems into different databases and on different platforms.

    The phrase “size matters” applies to believability and implementation success.

    Given current technologies, small to medium size companies find this “single copy” concept easier to implement. As company size and transactional volumes increase, the implementation becomes more difficult leading to higher implementation risk. If you doubt this, build and try a database with a few records. It works fine. To experience the other side of the coin, then build a database with several billion records servicing 100’s of millions of ERP transactions per day. Success lives in there somewhere, certainly not in the second case.

    To make “single copy theory work”, the IT consumers need to press the ERP companies to lead the charge. The ERP companies will need to form close relationships with database vendors to build, package and support their sales. The efficiencies enjoyed by the companies implementing “single copy theory” will contribute to, if not completely justify, the business case supporting the required technology refresh.

    This too is an opportunity created by the SAP Sybase partnership, both for the vendors and the IT community.

    1. Jim, I have to say I think you’re on the wrong side of history. The “one database for all” model is fraying at the edges, as far as I’m concerned; that’s why there are so many new entrants in the ADBMS market, the NoSQL folks talking about a non-database approach to managing some classes of data, streaming folks tackling an event-based look at data n motion, and so on.

  3. Merv,

    You offer a compelling case by mentioning DBMS history. One can not forget the huge “integrated database” projects of the late 70’s and their promise. One also can not ignore the efficiency of having fewer copies of data and the customers currently seeking that advantage.

    My view is that if enough customers are asking for a technology, then vendors should be looking at the business case.

    I am watching the “single copy” theory to see if it gains traction with customers. Yes, history can repeat itself and in doing so with modern technology offer different results. The combination of ERP and DBMS products can deliver impressive efficiencies. It is all about vendor engineering teams understanding size, volume, workload and putting it all together in a package. As of now the state-of-the art concept is to let the client integration teams do this heavy lifting.

    As to the new database concepts you mention, I look forward to seeing those implementations and the business cases reporting their success.

Leave a Reply

%d bloggers like this: