PureSystems: IBM Takes the Future by Pulling From Its Past


One of the aspects of the IBM I grew up with that largely was lost as the company struggled to compete on price over the last couple of decades was the intimate relationship the firm had with its customers. It was beyond just a vendor/client relationship at one time; IBM was imbedded with the companies and each large client had deep, dedicated subject experts who tied the company and its customers together at the hip.


This model was largely funded by a revenue stream tied to leasing IBM hardware and software. When IBM started to compete head to head on ever-more aggressively priced competitive bids, this intimate relationship generally suffered and much of the subject expertise, at least at an individual company level, was lost and IBM's customer satisfaction scores and loyalty suffered.


This week, IBM announced a program called PureSystems that represents a major enhancement to its recently announced appliance-based Expert Integrated Systems initiative. It's here that the company is recreating this tight intimacy but using technology that will allow it to scale while still remaining competitive on price. It is a really interesting move and one that only a company that has reached a steady state could implement.


Let's explore this fascinating concept of modularized expert systems integrated into cloud solutions.

The Value of New Technology


At the core of every argument we have ever heard from a technology firm is the idea that newer is better and that newer can be more economical. This is often tied to metrics that showcase how newer hardware can get things done much faster, newer software can be more easily managed and less labor-intensive, and newer systems can be more secure and reliable.


There is some truth to these promises because the development process takes the problems and unmet needs of customers and generally uses them to formulate the future roadmap. Products in short are better, or so we are told, because they are designed to be.


The Problem of New Technology


The learning curve is a bitch and no one wants to be at the bleeding edge of this stuff for while it can correct some problems, new technology often comes with a bunch of new ones with the difference being no one yet knows how to deal with them. This is where the subject experts who used to be intimate with vertical markets and individual companies used to come in. They either were trained on how to address the problems or they would warn their clients that the product wasn't yet ready for their unique environment. More important, if there were a problem, they could pull from peers and escalate to get the problem addressed quickly. This allowed firms that wanted or needed to be on the bleeding edge to continue to function.


IBM Expert Integrated Systems


Recreating this kind of service electronically is at the core of IBM's Expert Integrated Systems initiative. It places operational expertise in the systems, which allows administrators to hit the ground running and should shorten implementation time and increase results significantly. Rather than the typical list of soon-to-be-unmet promises, IBM has structured this modularized expert system in a way that the system itself assures the promises that sales made for it.


In concept this should lead to less learning on the job by the administrators because the systems know what to do, as well as better collaboration between sites virtually as experiences update the modules over time, and a faster ramp to promised benefits. Once implemented, the expertise is available throughout the lifecycle and the knowledge base on the company and its unique needs should expand dramatically, increasingly assuring future deployments and returning much of that earlier intimacy but using a mechanism that is more affordable and scalable than those earlier human-based efforts.


This is automated expertise, which, unlike the old subject experts, can scale with sales because it is built into systems like its PureFlex expert cloud system and PureApplication platform system. You own it and you help train it, but initially it is created as a gestalt of what came before. IBM is promising up to 300 percent increase in the performance of critical applications, 2x improvement in system utilization, 70 percent in cores reducing software licensing cost (and pissing off companies like Oracle), 50 percent lower management cost, and a 2x improvement in fault diagnostics over earlier systems. This didn't come easy and it was the collected effort of 37 labs over 3 years in 17 countries with a collective cost of around $2 billion.


When taken across a system, IBM reports even stronger results, twice the application density, 50 percent improvement in administrator productivity (it frees up their time), and up to a 75 percent lower system cost for the same level of performance. This is all the result of the built-in expert optimization function.


The solution includes servers, storage and networking in the same rack and all software is preinstalled and custom-optimized at the factory. Effectively, they are scalable appliances.


The underlying point is that the company that brought us Watson just announced a next-generation hardware and software system with built-in modularized expertise. For instance, if you install any IBM software product you will also install an expert module to automate it; this will eventually work for many third-party applications as well.


Wrapping Up: The Road to Self-Managed Systems


The days when IBM, or anyone else, can drop a bunch of subject experts into an account on a full-time basis are over. There aren't enough folks and companies won't bear the cost; however, getting the systems to a point where they contain this expertise is a long way down the path to full IT automation and a time when you'll be able to largely plug in and forget many of these systems, which will improve, just like a human employee, over time.


IBM may be the only company in the world that can get something like this to work near term and this is clearly a leadership effort. This is also the beginning of a path to put intelligence into systems so they can make their own decisions - in a few decades systems like this may be training us. The new IBM is looking a lot like the good-old Big Blue and it is getting there using its own technology. IBM feels so strongly about this that it will install the system, service it and provide lab support for a limited time at no charge. It is nice to see a company step up and it made this old IBM-er feel a bit of pride. That's just nice work.