Should we get out of the infrastructure business?

  • Vote This Post

    1
Posted on March 24, 2010 by Michel Labelle Posted as Past Post Archives
Computer Trash

Computer Trash

Brian Madden posed an interesting question on his blog.  Five years from now, will a 200-person company need any servers on site?

This is an interesting debate, but I would like to look one step further.  In a world where we can literally have everything in the cloud, do we really need to keep investing in our basic corporate IT infrastructure?  Is IT still a competitive differentiator?

Both Google and Microsoft have fantastic offerings that allow you to divest yourself of the traditional back-end servers like e-mail, document storage, collaboration and the like.  As more Software-as-a-service (SAAS) vendors come online like salesforce.com, we see the elimination of the need for CRM, ERP and other systems.  PRIMUS, TELUS and BELL offer VoIP in the cloud offerings for telephony, or alternatively some organizations are eliminating desk-side phones and going purely cellular.

The question becomes what is left?  As an organization that is growing internationally, we are already facing the debate over renewing our existing data centres at a local, regional or national level.  This debate isn’t about money anymore, it is literally about comfort zones.  By hosting our sites at a national collocation service I gain in reliability of the servers and eliminate the power, cooling and basic infrastructure maintenance from my already challenged IT budget.  But the debate shouldn’t end there.  Since 100% of our server are already virtualized, the natural progression will be to look at a pure cloud or hybrid private cloud.  Using a service like Amazon EC2 it wouldn’t be much of a stretch to say that I could essentially eliminate the data centre mentality/need from my organization completely.

I think true debate for the SMB crowd is about “roll your own” infrastructure or true SAAS .   There will always be monetary arguments both ways, but it comes down to the pace of innovation I believe.

With a roll your own, you will always be slower to deploy and innovate.  Basic change management comes into play here.  However you will end up with a tailored solution for your organization and that may be a competitive differentiator, or it may be an anchor around your neck.

With the SAAS, you will be on someone else’s timeline for innovation, however you will be innovating constantly as the products are updated and new features introduced.  The downside for your organization is the lack of differentiators between yourself and your competition.

I have my own views on IT as a competitive differentiator.  For me the commoditization of IT infrastructure has already occurred and as IT leaders we need to get used to that idea quickly.  That means that basic IT infrastructure on its own does not generate revenue in most organizations today.  The value we can derive from technology is focused on how quickly we can deploy it, thus allowing the rest of the organization to move forward quickly.  The days of spending months integrating products should be long behind us.

So to answer my own questions above;

  • No I don’t think a 200 person company will need their own servers.
  • No I don’t think most SMB organizations should continue to invest in their basic IT infrastructure but should look at outsourcing/SAAS instead.
  • And No, I don’t think IT is a competitive differentiator for most SMB businesses, but is more of a drain on resources and talent.
Michel Labelle Michel Labelle (31 Posts)

Michel Labelle is IT director for Global Container Terminals. He has over 20 years of industry and consulting experience at various organizations. Prior to this position, Labelle served as director of infrastructure technology with TSI Terminal Systems Inc, where he was responsible for the IT operations, infrastructure and support teams. His duties included the ongoing management of server, network, and voice and data communication activities. With a background in IT risk management and IT forensics or multiple organizations, including Auditor General of Canada and Transport Canada, Labelle brings a rounded approach to IT operations. Labelle has served on the federal Canadian Critical Infrastructure Protection Committee and is active in promoting digital civil liberties. A regular contributor to IT World Canada and other industry publications, he has written for MacWorld, IT Architect, Network World and Network Magazine, and has been a featured speaker at VMWorld, IP4IT and the MISA conference. He trained at Heritage College in computer science with qualifications in project management, systems design, and voice/data communications systems deployment.


No Related Posts
Tags:
  • Samy Benzekry

    You know the question you ask is very close to what Nicholas Carr (http://www.nicholasgcarr.com/) asks in his books, does IT matter and The big Switch. The essence of his thinking being, in the beginning of the century the competitive differentiator was do we have enough power to power our plants. At first, the plants where powered by flowing water thru a water wheel and via an intricate system of cables, would move the production equipment. Then came the use of electricity in the manufacturing process. The early adopters actually used the water to manufacture electricity on site, a few years later, every major competitor had the capacity to manufacture electricity and had transformed their production line from using mechanical machines to electrical machines. Essentially, the huge costs associated with creating a constant supply of electricity was no longer a competitive differentiator, it was down to how good where you at doing it and how cheaply can you produce it. With the advent of the electrical grid, the early adopters of the grid could and still do to this day enjoy electricity at a fix or negotiated price. Most importantly, in almost in every case that cost was lower than manufacturing it on site. So, when the manufacturing electricity is no longer a competitive differentiator and is accessible to everyone for almost the same cost it becomes a commodity. And in economics, when something is a commodity, it becomes a game of where can I get it at the cheapest cost. I won’t go there but think of what happened in the telephone industry post deregulation. The intelligent corporations understood early and connected to the nascent grid in the early days of the previous century. The capital that was freed from this where re-invested in manufacturing better product or simply passed on to the consumer in the form of cheaper products. Is the cloud computing paradigm akin the electrical grid? I think we are getting closer of a pay as you need computing needs, but, like my kids like to say, “are we there yet” (with wide usage of cloud computing)? For the most part, I do not think so but by far, the fastest growing segment of real life production usage of cloud services is SAAS. Salesforce has been around 10 years and boasts 72,500 clients (http://www.salesforce.com/company/), Human Factors (HR type of SAAS offering) managed to wheel in Siemens. Personally, I think we are living in very exciting times where we are witnessing the start of fundamental change in the way business processes, supported by an IT infrastructure, are being deployed on .