Blog post

Is Cloud Computing Killing Open Source in Government?

By Andrea Di Maio | September 02, 2009 | 15 Comments


I know that some reader will raise their eyebrows reading the title. After all, cloud computing is intimately connected to open source: Linux servers are at the basis of most cloud infrastructures, and several applications that can be consumed as a cloud-based service are either open source or based on open source components. One may be almost tempted to look at the two as being strictly intertwined and mutually dependent.

Well, maybe they are from a vendor perspective, but a client conversation this morning confirmed my suspicion that they may be on a crash course, as I already wrote a few months ago. I was chatting with a CIO from a local authority in the UK and we were discussing about the outlook of open source software deployment in local government around the world. The UK government published an interesting open source policy (a Gartner research note for subscribers is available here), but it appears that not much happened since its publication. A recent survey of local authorities showed that many are still waiting for more guidance around product assessment and maturity models from central government.

But the new buzzword, both in London and in Washington, is cloud computing, and in particular the definition (and – maybe – development) of a government cloud (nicknamed G-cloud). So open source software is certainly losing momentum and political appeal, while cloud computing is gaining press coverage and executive interest.

More concretely though, and this was the point of our conversation, cloud-based alternatives to proprietary infrastructures, operating systems, office productivity suites, web servers and other applications, are becoming more palatable than open source software. First of all, going open source does not free a government agency from a vendor, who will provide an open source product (usually cheaper than a proprietary one): government agencies downloading their own operating systems or word processors from community sites and maintaining them with internal resources (including contractors) are in a tiny and rapidly declining minority.

Therefore some of the primary drivers to choose open source, i.e. cost and vendor independence, are just going away: in most cases cloud-based solution are going to be cheaper (and more elastic), and to use open source one has to go through a vendor anyhow. As a consequence I have seen a drop of interest in open source and corresponding surge of interest in cloud computing to solve pretty much the same problems (how do I reduce my dependence on Microsoft? how do I save on licensing costs?).

Of course there are still plenty of reasons supporting the importance of open source.

  • What if you have to change vendor (because it goes bust or just becomes too expensive)? With open source there is a chance that another vendor or at least a community can support you, whereas portability between cloud services is still far from reality.
  • What about the impact on local economy? Many have predicted the beauty of open source in government on the basis that it would create and sustain a local ecosystem of IT skills. This is not the case with cloud-based applications, that most likely run somewhere else without using any of the physical or intellectual resources in your jurisdiction, and without even indirectly benefiting your economy.

However these issues pale if compared with increasing budget constraints, hiring freezes that prevent from refreshing skills, financial vulnerability of small vendors and so forth. Cloud computing implies risks, but creates economies of scale that can benefit large as well as small government organizations.

Let’s face it. While open source software has often increased the need for application development ad other technical resources, cloud computing may ultimately make IT go away. Which one do you think sound more appealing to government executives?

The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.

Comments are closed


  • Phil Mocek says:

    Cloud computing and “open source” (which you seem to use to refer to [free software][1]) are apples and oranges. A better comparison can be made between hosted software services, a situation in which someone runs software on his own servers for clients, and running software on one’s own equipment.

    I’m curious what you meant by “cloud-based alternatives to proprietary infrastructures, operating systems, office productivity suites, web servers and other applications”. Would that include anything at all that runs “in the cloud”? It seems that alternatives to proprietary operating systems and software are open or free (as in “free speech” not “free beer”) software and operating systems.

    You wrote that “going open source does not free a government agency from a vendor, who will provide an open source product (usually cheaper than a proprietary one): government agencies downloading their own operating systems or word processors from community sites and maintaining them with internal resources (including contractors) are in a tiny and rapidly declining minority.” You don’t need a vendor for free software. Typically, there *isn’t* any vendor. It just exists. It’s part of our public body of knowledge, and you can use it for any purpose you like without asking permission from anyone.

    Regardless of the license under which a piece of software is available, users might transfer it onto their own computers across a network or on a piece of removable storage media. Referring to the use of standards-based free software as “downloading their own operating systems or word processors from community sites” sounds like FUD from companies who sell software. “Downloading” may have some negative connotation with those who are under the spell of the MPAA, RIAA, and similar mafioso, but it simply means to transfer bits from one computer to another across a network, and it’s the way most information is moved from one place to another nowadays.

    It’s extremely short-sighted to focus on the gratis nature of free software while ignoring the libre nature of it.



  • Phil Mocek says:

    The link I attempted to leave that was apparently scrubbed is the Free Software Foundation’s “What is free software and why is it so important for society?” at .

    For those who are unfamiliar with the concept, I’ll quote FSF: “Free software” is a matter of liberty, not price. To understand the concept, you should think of “free” as in “free speech,” not as in “free beer.”

    Free software is a matter of the users’ freedom to run, copy, distribute, study, change and improve the software. More precisely, it means that the program’s users have the four essential freedoms:

    * The freedom to run the program, for any purpose (freedom 0).
    * The freedom to study how the program works, and change it to make it do what you wish (freedom 1). Access to the source code is a precondition for this.
    * The freedom to redistribute copies so you can help your neighbor (freedom 2).
    * The freedom to improve the program, and release your improvements (and modified versions in general) to the public, so that the whole community benefits (freedom 3). Access to the source code is a precondition for this.

  • kaiyzen says:

    What would be some of the open source solutions that would be abandoned for cloud solutions?

    If gov’t is trying to descrease dependency on MS and saving $$ you can look at MS’s core product lines, Windows, Exchange, Office.

    In moving to a cloud based solution what are the options that gov’t is looking into?

    WIndows aside.., for office and exchange replacements you have Google products, zoho, few other providers othere there, but is gov’t really interested in having their data site on the outside?

  • ben breeland says:

    The government’s implementation of cloud computing is to create shared datacenters that use some of the features of cloud computing to give government users a private cloud. With this type of implementation, the government simply deploys existing hardware and software centrally and does not take advantage of the service nature of cloud computing. Cloud computing provides unlimited access to a range of services where the consumer only pays for the services used. In cloud computing the request for an email service does not specify Microsoft Exchange. Instead, the specification is for five email accounts accessible from the web. The vendor offering the service chooses the email software (open source or commercial) that meets the needs as well as provides for always on, scalable, and measureable. Cloud computing suggests that it is all about the service and not the components. The government has to move to consider a move to a service-oriented architecture and focus on the specification (measured with SLAs) rather than the components. No one repairs or troubleshoots a phone when it fails; instead, a new instrument replaces the faulty one. Restoring the service happens before repairing the component. This is cloud computing – an open source component is just as good as a commercial one to the cloud computing vendor.

  • @Phil – thanks for your comments.
    I do not think I ever alluded to “open” meaning “free”. If you are familiar with Gartner (and my own) research on open source software, you should know that this is a confusion that some of our client were making in a distant past, and – luckily enough – does not seem to occur any longer, but we have also had the distinction very clear.
    Having been following OSS in government for quite some time, I can tell you for sure that there are only few cases where government agencies download open source software and maintain it themselves (i.e. with internal resources. In general open source gets into a government organization either through a supported, licensed product (e.g. RedHat, Novell, Sun, etc), or through a system integrator (which can use open source components to build or run a system) or through a service provider that supports it.
    Also, this is the first time I hear that “downloading” has necessarily a negativa connotation: today most software gets downloaded both for consumer and enterprise use (interestingly enough, in certain regions like Europe that makes it a service rather than a product – as it would be if delivered on a physical medium – and saves software providers by product liability laws).
    The bottom line is that software that really is a commodity for government organization should have the lowest possible TCO: I would argue that cloud-based apps are better positioned for this than open source applicatioms deployed internallyy.
    Of course there will be load of cloud-based open source applications accessed as a service: but then, why should clients care that they are open source ?

  • @ben – your final sentence summarizes very well what I meant and what I said in my earlier reply to Phil.
    You also touch on a very interesting aspect, which is the nature of a government “private cloud” and to what extent this is more a consolidation of existing government resources rather than a truly innovative sourcing model.
    You may wish to look at some of my previous posts, where I have postulated that cloud computing may become an enemy of government consolidation. What is quite interesting is how the perspective of central government organizations that run large infrastructures and position themselves as government cloud providers differs from the perspective of prospective client agencies that are looking at cloud-based solutions to reduce their capital spending and commoditize infrastructure (which means, they may well choose a vendor service over the central government agency one).

  • Phil Mocek says:

    Ben Breeland wrote, “an open source component is just as good as a commercial one to the cloud computing vendor.” This implies that software is either commercial or open source, but that is not the case. Commercial software is sometimes available as open source. Similarly, “free software” does not mean non-commercial.” Quoting the Free Software Foundation [1]: “A free program must be available for commercial use, commercial development, and commercial distribution. Commercial development of free software is no longer unusual; such free commercial software is very important. You may have paid money to get copies of free software, or you may have obtained copies at no charge. But regardless of how you got your copies, you always have the freedom to copy and change the software, even to sell copies.”

    Andrea: So when you referred to “open source software” did you mean to include all software — free and non-free, commercial and non-commercial — for which source code is made available?

    I am not familiar with your or Garner’s research. Could you please cite some examples of the publications of yours and theirs to which you referred in your response?

    If you intended to compare to the situation surrounding use of proprietary software to that of alternatives, and if by “open source software” you meant “free software”, then use of the latter *will* free government from a vendor, because in that situation, they are free to use the software however they see fit, without any vendor’s involvement. With non-free software, a vendor of some sort is a necessity.

    By writing that “cloud-based apps are better positioned for this than open source applications deployed internally” you again implied that applications are either open source or cloud-based but not both. These are completely different. Open source means that source code is available. Cloud-based means that software is run on a remote system and offered at a service. You’re still comparing apples to orange crates.

    You asked, “why should clients care that [cloud-based, open-source, applications] are open source?” Good question; I don’t know the answer. Freedom is much more important, I think, than availability of source code. The former is, in some cases, a prerequisite for the latter.



  • Mike Kavis says:

    I disagree with this assessment. One of the beautiful things about cloud computing is that processing power and disk are so cheap it is now feasible to build a lot more redundancy both within a virtual data center and across virtual data centers. I no longer have to be as conservative about how many servers I stock pile due to floor space, air conditioning, electricity, etc. It may be very strategic for me to be more distributed than before. Since I might have a large number of application and database servers distributed over many virtual datacenters, it is still critical that I don’t get overwhelmed with licensing costs. Choosing the LAMP stack over Windows saves me huge amounts of money when you add up the overall savings per virtual image per hour.

  • @Mike – you care about licenses if you look at infrastructure as a service. But if you look at software as a service or platform as a service, that distinction – i.e. proprietary vs. open source – goes away from a client persoective. I do pay for what I use, and I will choose according to value for money: whether the cloud service providers uses a proprietary product or not makes no difference to me, provided the price is fair.

  • @Phil – we have always used the OSI definition of open source software (see Our research is available to Gartner clients through there is plenty on open source (a good starting point could be a research collection about the “State of Open Source, 2008”, available at to subscribers).
    I do know that cloud based applications can be either proprietary or open source or a combination. The difference I was trying to explain was the one between (a) using a cloud-based application (SaaS) and (b) using an open source application deployed on premise (i.e. running on your own infrastructure). Of course there is a case (c) proprietary application deployed on premise: the total cost of ownership of (c) can be greater or lower than (b) depending on many factors, license costs being only one. Now, in my post I was addressing those who look at (b) as the way forward, and made the point that (a) may be much cheaper than both (b) and (c), hence making the open-source vs proprietary battle pretty much useless.

  • Gary Wilson says:

    You may be interersted to know that RedHat have just released deltacloud ( ) with an aim of offering a basic API that is standard and works against a number of different cloud providers. This is a in beginning in cloud integration services and other have come before and are out there. However, we are getting more clouds and whether deltacloud and others can keep up is another story. However, that said the emergence of an agnostic cloud framework is not too far off in real terms and will evolve with the cloud growth and pervasion.

    As Peter Nikolov of 3tera pointed out the other day ( ), this is a new paradigm and we are just starting to seeing this new technology take off.

    However, the cloud there is a lot of marketing bullshit around the cloud as well. The cloud as you point of outsources technical knowledge. This is important, because proper production cloud infrastructure can be even more complex than traditional infrastructure and it can grow and change quickly as well. Ironically, achieving zero single points of failure in the cloud are just as hard, if not more techincally complex than achieving zero SPFs in physical infrastructure. That is if one sees the cloud provider as a single point of failure, which because most, they are just that, geographical single point of failure complex platforms.

    In the past when we saw a failure effect one site, now we will see a failure effect a cloud and multiple sites. Therefore, you need full production appliications on 2 clouds, fully replicated full stack applications in two different clouds? Even though your entire stack on the cloud maybe have full redundancy, you cannot ensure that the cloud will have the same.

    The cloud has just taken a number of little failures that happen often and sunk them all into a system where they happen less. However, the clouds will break as well at some point. They already have, Amazon’s EC2 presence in Europe was taken offline during to a ligthning strike.

    Clouds are not going to go away soon, it is unlikely that any of things will stop or slow the rampant growth of the cloud concept as it makes sense. This is due to the fact that they make sense to support our informational growth. Fast scaling,”unlimited storage” and scalable bandwidth, instantly. The driver behind the cloud is not business based. Although it does have positive points for business. The has increase in demand for scale and instant on and growth is what is driving the cloud. It is driven by infromational evolutionary, in age where all data has value and is increasing in volume at an increasing rate.

    However all the marketing and hype is not saying there could be difficulties ahead.

    The cloud is a good business model and due to complexity and their different languages, they are going to be racking in money as we move our ever increasingly infrastructural and data storage demands onto them. We then get charged per GB to store our ever increasing data, per byte of transfer it when it is accessed and per CPU hour to analysis and work with it.

    In old world ISPs we used to use the term customer stickiness (and still do), e.g. how much pain is it for someone to move to another ISP taking into account all their mail addresses, hosting, dialup connectivity configuration, etc. I think customer stickiness has just been reinvented in the clouds.

    However, it is here to stay so we better get used to the new complexities that it brings with it.

    Onward to the agnostic cloud and then the single cloud, where all our own laptops and desktop machines are partly added to a distributed infrastuructural cloud, Boinc’ing, MapReduc’ing and P2P’ing the entire Internet as we start utilising the hardware which is in use more optimally.

    We do live in interesting times.

  • Andrea, no I don’t believe cloud will kill open source, quite the opposite. I know some will find it hard to believe but we can focus on more than one thing at a time, so just because we go quite on a few things do not mean they are off the agenda.

    In the Government Private Cloud, it will be a mixed economy. We have no choice as the European procurement law for the Public Sector limits our ability to specify brand or product – the vendors select the best product to fulfil the requirements. However we believe there is more flexibility around open source as by definition there is less of a procurement issue (open to all) and the creation of the app store creates visibility ( a key constraint of many software vendors) as well as a provisioning model which will ease implementation. European procurement lawyers your advice would be warmly welcomed on this!

    But look at it this way. If we have created an infrastructure as a service, and a platform as a service (or multiple) and we have software as a service as well as an app store, this is likely to play to both the SME market as well as the open source market.

    Often a constraint for some suppliers in these markets is the ability to provide scalability, robust infrastructures, security mechanisms as well as limited resources to market, sell and service etc. G-Cloud will have to “wrap” all applications in technology to protect the rest of the environment, provide test environments, and provide provisioning models etc. As G-Cloud will be IaaS and PaaS and SaaS purchase decisions will begin to focus on greatest capability (in all contexts) for the lowest price made available to the end consumer in the shortest time. There is no reason why open source will not shine in this environment.

    To confirm open source, open standards, reuse, standardised desktop models, Green IT, G-Cloud (with an app store), rationalised networks and data centres, Professionalisation of IT enabled business change, shared services, supplier management, information assurance and security, technical standards and architecture and reducing driving through efficiency etc etc etc are all still on our agenda.


  • @John – Thanks for your reply, you are always most welcome.

    I know that the UK government is committed to an aggressive transformational agenda, with many different facets from a technology and management perspective. Nor do I believe cloud computing will be a silver bullet solving many of the technology management challenges that government department CIOs have to deal with.
    As we said in our research, cloud computing is just an alternative sourcing model, which will coexist alongside make, buy, open-source, share & co-develop. What I have noticed in multiple client inquiries is that attention has shifted toward off-premise, on-demand approaches as a way to decrease cost and increase flexbility. Of course there will still be internal deployments where the use of open source solutions or components (either directy or through system integrators) will be an option: in this respect the UK open source policy is quite well structured to help departments deal with the most important questions.

  • Open source is alive well in the government, even with some of their own. TVA, a government entity, has released the openPDC (, a tool for high-speed data collection related to the transmission side of the smart-grid as “open source”.

  • @J.Ritchie – Thanks, this is an excellent example of where open source still makes a lot of sense, i.e. to solve a specific vertical problem