Hoov's Musings

How can you be in two places at once?

            “…when you’re not anywhere at all?”

By Dave Danielson (Guest Muser)

This title from one of my favorite comedy albums from the early 70s came to mind as the appropriate way to summarize my thoughts and opinions about the course of application management over the last few years.  (Of course, I also have to acknowledge that if I am referencing the 70s, maybe Mark is right and I really am getting old).  My other potential title for this musing was to quote Bill Murray’s famous line “I’m a schizophrenic, and so am I” – but it’s probably better not to go there.

Those of you who have been readers of the Musings over time know that I have written about application quality management (AQM) in the past and more recently I have had the opportunity to “live it” as the interim CEO of Altaworks – a company that provides leading edge distributed application monitoring, correlation and analysis.  This has given me a chance to  observe first hand how customers are trying to solve application performance and management problems, to evaluate how vendors are and are not rising to this opportunity, and to test out the theories and suggested solutions that I have been arguing for.  You might say that I’ve been conducting a survey of customers over the past 18 months where “voting with their wallet” is the only way to keep score. 

So, what did I learn from this experience?  I’m feeling conflicted because I was right about the problem statement and the required solution set but I missed the timing and the customer’s inclination to err on the side of “safe” vendors even at the expense of the right solution.  The all-too-often result is that a startup in this space can have the right product at the right time, but can still find itself “nowhere”.  Let me explain…

By now, it should be very clear that the application architecture of choice is Java-based applications, implementing WebSphere or WebLogic at the application tier and with well defined N-tier architectures.  Also, with load balancing technology being commonly used, these applications are deployed across an array of servers from 10 to 100.  When these mission-critical applications fail to perform there are three basic questions that need to be answered:

• What is the application doing?
• What’s behaving abnormally?
• What transactions are being impacted?

For more than four years Acuitive has been writing about and presenting solutions for these problems under the general heading of “Application Quality Management”.  Also, throughout most of this decade a number of startups have been working to provide solutions for these problems.  These solutions are generally grouped by the way they address one of these basic questions.  I place them in the following categories:

·        Application topology and dependency mapping – Companies like Collation, Relicore, Troux, Rendition Networks, and others.  These products discover and display a “map” of the connections and component dependencies.  These displays are very useful to understand all of the distributed “pieces” of the application.

·        Performance Monitoring, Correlation and Analysis – Companies like Altaworks, Dirig, Panacya, Cyanea, Performant, Quest, Wily, Precise and others.  These products monitor detailed metrics and uncover correlations of behavior.

·        Transaction Tracing – Companies like Openview Transaction Analyzer, Appilog, Tonic Software, and some partial tracing from developer tools (Quest, Wily, Compuware, etc.).  These products follow a single transaction through from client to varying levels of “depth” into the application infrastructure.

My prediction would have been that after the market began to recover in mid-2003, the surviving AQM startups of all types would have had some pretty good success selling to large enterprises.  But that hasn’t happened.  Why?

Most large enterprises would agree that their business is critically dependent on multi-tiered applications.  It’s pretty typical to find a large enterprise with 100-200 Java-based distributed applications with multiple tens of millions invested in development and support.   But large enterprises have reacted in a very cautious way to AQM startups.  Enterprises are willing to spend money, but a key issue is that they now demand more of a “whole product”, or complete solution. Start-ups who have developed one aspect of AQM deeply are less preferred than those who touch all parts.  Instead, large network management vendors that have account control like HP, BMC, and Tivoli have remained strong even though they don’t have the leading-edge products.  But they do have the marketing muscle and customer influence required to tell the customers “hold on, we’ll have that out real soon…”  But customers can only hold out so long. 

Filling the void,  there are some newer AQM companies with reasonably broad “solutions” established prior to 2004 have been deemed by enterprises to be stable and safe enough to work with as well.  The list includes Mercury, of course, Quest Software, Veritas with their acquisition of Precise, and Wily Technologies, who have been growing strong recently.  These vendors are calling attention to the importance of application management and taking market share away from the established network management vendors. But they are also making it almost impossible for an AQM “component” vendor to get any traction at all.  The lesson is that if you achieved critical mass before the recovery, you could be a contender, but you can’t expect to draft the recovery to become one for the first time.

The large management software vendors – BMC, HP, Tivoli, and CA – as well as many mid-tier vendors – Concord, Smarts, Micromuse, NetIQ, and others – are all seeing the success of Mercury and Quest and are actively trying to put together a complete solutions of their own.  Mercury, for their part, seems to be convinced that if they acquire enough companies, they will have a “check-the-box” for any and all RFPs.  In doing so, they are creating a lot of activity (both rational and irrational) by their competitors who are afraid that they can’t keep pace or satisfy their customer base.   Actually, it’s more like most traditional Enterprise System Management (ESM) software providers are living in fear of Mercury while the developer software companies like Rational and Borland are more worried about Quest.  This speaks to the more fundamental issue of how does application management get done in the final analysis.  If it plays out as an extension of the application lifecycle then the companies like Borland, Rational and Quest are in a better position to win.  If, on the other hand, the solution will come as an extension “up the stack” from an network management vendor, then Mercury, BMC, HP, Tivoli and CA will be in a good position to pull together the complete solution.  I, for one, am voting for the lifecycle approach and “designing in” performance management into applications, but that’s the subject of another musing.

All this, and the mindset of over-extended investors of a number of the smaller application management providers who are still experiencing market tractions problems, has led to a merger/acquisition frenzy of sorts. 

Large enterprises for their part are still primarily driven by strong ROI and OPEX-reduction arguments. They have shifted much of their attention away from AQM to data center configuration and change management products.  Since there are few (if any) large, safe incumbents in this space, emerging vendors have performed better.  Companies like BladeLogic, mValent, ConfigureSoft, Opsware, Evident, Ejacent, and others are offering strong solutions for simplifying change management with software application configurations over large arrays of servers.

Another thread of interest in the enterprise is around Service Level Management – offering a way of cost-justifying IT budgets by proving service value to the company.  Startups like Appilog, Managed Objects, Proxima Technologies, Oblicore, Opticom, and many others are making it possible to monitor service levels that combine network, application, security and legal/contract information.  These products are starting to bring some reality to the oft-discussed Business Process/Activity/Service Management (BPM, BAM, BSM) – a key objective of large enterprises. 

When you sit back and try to “take score” as I do from time to time, it becomes clear that the problems that have been faced by AQM  start-ups are quite understandable.  The bottom line is that at the same time the  market was recovering enterprises seeking more complete solutions. The vendors that failed to see this evolution became “features” and not “products”. The vendors that either were fortunate enough to ramp up early or who saw the product evolution and correctly anticipated the change were the ones to thrive.  The situation for those less fortunate is that they survived one of the toughest markets in recent memory only to have to figure out how to sell or merge their company with a larger player in the market.

As a result, it’s “open season” on small startups that survived the market and thought they would have their time.   The next 12-18 months will continue to see rapid consolidation and some of the names that we know today will be gone – big names as well as small.  So as I sat to write about this topic, I felt torn between being right about the technical solution but “off target” about the inevitable dominance of large vendors and the way that this market would play out.  The good news, however, is that those companies that take action now – to be a part of the complete solution required by large enterprise, can still be winners. 

That’s the approach we’ve taken at Altaworks.  As the interim CEO for Altaworks, I have lived through this last 2 years and had to come to the realization first hand that the optimum solution was to move toward finding the right acquirer.  I am pleased to announce that Altaworks has been acquired by Opnet Technologies – a company that I am convinced is well position to cover many of the needs of application planning and management for large enterprises in the coming few years.  I’ll talk about this more next month – when I hope to be more clearly in one place on the issue of Application Management.