Note: This is an edited version of an analysis I wrote as a Council Member with Gerson Lehrman Group (http://www.glgroup.com) and posted to GLG members on August 28, 2007.
The high cost and relatively inflexibility of information technology is frustrating for many business managers. My experience running large technology infrastructure organizations for investment banking and financial services firms taught me this first hand. Why does it cost $1000 a year or more just to support a desktop, not counting the cost of hardware and application software? Why can't ease of use and flexibility be as good as many people now have at home, with a standard PC and an internet broadband connection? These questions will be familiar to almost any IT support manager!
To protect customer data, provide information security and manage costs, IT shops must typically lock-down desktop PC and server standards, denying flexibility and rapid deployment to application developers. At the same time, the demands for data storage and archiving, high reliability, and continual cost reduction seem insatiable. Unfortunately, the traditional client/server model seemed ever more poorly equipped to deliver the necessary results. Unfortunately, there were few alternatives.
Thin-client solutions had limited applicability and didn't really save all that much money. Solutions, such as Citrix, worked reasonably well for a small number of well defined desktop applications distributed across a large branch network. However, the cost of managing and maintaining the back-end and related networks was often significant. Additionally, this solution did not scale to application-intensive environments such as a securities trading floor or a large and complex head office.
Alternatives such as Clearcube's system, which puts blade PC's in a server room off the trading floor were one solution, but the price point was very nearly identical to traditional client PCs, with the added complexity of concentrating the power load and heat dissipation requirements into that very same server room, which anyway could not be far from the end users.
Nevertheless, the direction is clear: traditional fat-client desktop installations are no longer sustainable. It is simply too expensive to manage large desktop installations, particularly when a desktop visit by tech support is required for certain types of repair tasks. Furthermore, desktop compute power cannot be shared by multiple users. A power user needs a high-end desktop, even if it sits idle overnight and on weekends. Solutions for reliable thin-client technologies are maturing, and the product by Pano Logic described in the attached weblink are yet another example of this trend.
The coming paradigm shift was further brought home to me while working on the technology for my current consulting practice. After over a decade in large financial firms, I had perhaps gotten used to an IT price point in the tens of thousands of dollars annually for some users. Needless to say, a boutique consultancy does not have that kind of money. More importantly, we didn't need it. A few low-end Apple computers, open-source software, and widely available networking and e-mail solutions costing a few dollars a month mean my price point per desktop is measured in tens of dollars a month, not hundreds. This is not a mere 10% or 20% savings - this is an order of magnitude lower cost, with nearly all the same usability and security benefits of large, legacy installations that I had been managing. This is even before considering alternative desktop application solutions over the network such as that now being offered by Google.
So are alternative solutions ready for prime time? I believe that increasingly the answer is yes. First of all, we've already solved the networking problem: The standard is now the IP protocol running over ethernet or WiFi local area networks. No longer do we deal with SNA, DecNet, IPX, or a range of other protocols that were common place just a decade ago. Next we need to tackle the desktop, data storage and management, and mobility requirements.
IT managers will need to review the business applications currently in use. It remains the case that many legacy applications are difficult to port to thin-client solutions. By the same token, many of them are probably not really needed. The decision-making process to retire such applications requires close business/IT cooperation, but can deliver significant rewards. Perhaps ironically, Sarbanes-Oxley (and its equivalent requirements around the world) actually helps here: Since business managers are required to understand their business processes, and by extension the applications that support them, the excercise to cull unnecessary software should already be underway!
In addition to desktops and applications is the complexity of data management. I would hazard a guess that much of what people store is unnecessary e-mail with too many attachments. Thus, the next piece of the paradigm-shift puzzle is data storage. Here the problem is trickier, but the bottom line is that it is simply too hard to find things when you need them. This is true both for the individual employee trying to dig out that Powerpoint file from 3 months ago as well as for the Legal department manager who needs to track down particular e-mails to comply with a court-order for documents relating to pending litigation.
The trend is moving beyond mere data storage, to data management. This will take several forms: (1) Application managers will of necessity get better at imposing data architectures to which all applications, in-house and off the shelf, must adhere. (2) Rules for tagging and archiving e-mails will get more sophisticated, so that necessary e-mails are kept and logged, and unnecessary ones get deleted after a defined period of time. (3) All other application files will be similarly tagged and culled or archived on defined schedules. (4) Tape back-up will not be used for archiving data that is regularly needed.
There was a time when technology infrastructure managers simply had to deliver storage solutions with tape back-ups. This will change, as infrastructure will increasingly be bundled with applications software to manage data content.
Lastly, the workforce will become increasingly mobile. Fortunately, efforts to centralize the back-end and improve information mangement will make mobile computing easier. A networked worker will be able to access everything from nearly anywhere, with security controls varying depending on whether the access point is in a company office or over a remote network. The holy grail of free-seating and mobile workforce are increasingly in reach.
Enterprise computing is shifting to a new paradigm: Thin, Lean, and Mobile: true thin-client desktops, lean application suites with even leaner data management, and increasingly mobile access including free-seating, WiFi, and mobile wireless.
Firms that deliver cost-effective, reliable, and scalable solutions that fit this paradigm will be the winners in the enterprise technology vendor market.
