Linux option $100k more than Windows,
claims Members Equity Bank


news Small Australian bank Members Equity Bank (ME Bank) has issued a joint statement with Microsoft claiming that using Linux as its core operating system underpinning its new core banking platform would cost $100,000 more than using a platform based on Windows Server 2012, but without providing any evidence for its claim.

In a statement issued overnight by Microsoft, ME Bank, which is a small bank based in Melbourne, noted that historically it had depended on two core banking systems — a Windows–SQL-based product called NTBS that managed all savings and deposits; and a Solaris-based product called ULTRACS for loans and transactions products.

“Updating our legacy banking systems was becoming increasingly difficult,” said Jem Richards, Enterprise Architect, ME Bank. “Any level of change required custom development, which meant bringing in specialists from outside. This cost us tens of thousands of dollars a year, caused delays and stymied innovation.”

According to the bank, its IT staff were eager to support business goals and reclaim their agility and ability to deliver turnkey solutions in a timely manner. This led to the search for a single, integrated core banking system. “The IT environment is diverse and constantly changing and it was challenging for our team to keep up as they were always immersed in tactical and reactive tasks. We realised that if we consolidated our IT skills on one technology, we could build our in-house expertise to enable us to make strategic changes more easily and quickly,” said Richards.

Consequently, in late 2011, according to the statement, ME Bank selected a new, platform-agnostic banking software system, Temenos T24, to manage all its core banking operations. It then began a review process to look at which operating system would be the most cost effective for its needs. There were two equally viable options to deploy it on Windows Server 2012, with the SQL Server 2012 database, or a Linux-based solution. The statement did not note which database software would be used on the Linux platform; the most common options are usually open source solutions such as MySQL or proprietary options such as Oracle’s database packages, or IBM’s DB2 solution.

“We conducted extremely rigorous comparisons on costs, the availability of technical support and scope for in-house IT skills consolidation. Although one operating system option was essentially free to deploy, based on our past experience, we knew that it would cost more to support than Windows. In addition, Microsoft and Windows community specialists were readily available to help us configure Temenos T24, whereas finding the relevant skills for the alternative platform with Temenos was proving to be a lot more difficult, “said Richards.
ME Bank said that it found the running costs of a Microsoft database versus the competition also differed substantially, which would affect its future growth plans.
“At present growth rates, we would soon need to upgrade the alternate platform solution, which would cost over $100,000 more,” said Richards. “In contrast, our existing Microsoft licence will allow us to grow without additional cost. When we analysed all the costs over a five year period, it was clear that if we deployed our core banking systems on Windows-SQL, the total cost of ownership would be substantially less than if we chose the alternative.”
ME Bank made the decision to purchase Insight, Temenos’ business intelligence software, which used SQL-based reporting and analysis tools. This, according to Microsoft’s statement, has allowed the team to grow their IT skills and capabilities on one technology to ensure increased agility which has enabled them to make changes to their environment themselves and introduce new customer services faster.
“Selecting Windows–SQL enabled us to consolidate internal IT skills on a single technology, and build our in-house capabilities in line with the business,” said Richards. “Ultimately, this means we are more in control of our banking platform and reporting capabilities.”
By deploying a new core banking system using Windows and SQL Server, ME Bank said it could “avoid major future upgrade costs”, and “consolidate IT skills in a way that promises maximum scope for agility in the future”. It said it has decommissioned a mixed Windows-Solaris environment and consolidated its core banking system to one technology, reducing the dependence on external consultants, boosting in-house capabilities, and saving tens of thousands of dollars each year.
According to Microsoft’s statement, ME Bank is also confident that deploying Temenos T24 on the Windows Server platform will maximise development choices over the life of the core banking system.  In addition, the bank believes the all-Windows platform should give it greater flexibility. “With the core banking system on an all-Windows platform, we will be able to standardise the infrastructure more easily. This means we will encounter fewer issues if and when we move those systems to the cloud,” said Richards.

Look, I’m not going to go all-out here and claim that ME Bank’s view, as stated in Microsoft’s press release, is complete bullshit. The bank wouldn’t be putting its name to this kind of statement if it wasn’t willing to back up its (somewhat audacious) claims.

However, there are a few things here which do appear a little dubious, and which I’ll ask ME Bank if I can organise an interview with them. For starters, we’re not precisely comparing apples with apples here. Microsoft is providing Windows Server and SQL Server here, but it’s extremely suspicious that this media release doesn’t mention either which variant of Linux ME Bank evaluated (Red Hat? SUSE? Ubuntu? Slackware?) or which database it evaluated. I’d be extremely surprised to see something like MySQL used in this kind of core banking environment — it’d be more normal to see an Oracle or an IBM database here. Could this account for some of the cost?

Secondly, this sentence just doesn’t make any sense: “At present growth rates, we would soon need to upgrade the alternate platform solution, which would cost over $100,000 more,” said Richards.

Is ME Bank really saying there is a cost difference in upgrading a Linux platform which you don’t see with a similar Microsoft platform? In my experience it’s usually the opposite — Linux licences come cheap and the platform tends to make great use of available server hardware, whereas Microsoft’s software, up until a few years ago, was known for being more of a server hog. Plus, Microsoft licences don’t precisely come cheap … a few copies of Windows Server 2012 may be available through ME Bank’s existing Microsoft enterprise agreement, but it’s common knowledge that those agreements are often re-negotiated on an annual basis. Microsoft doesn’t give away anything for free, and I wouldn’t be surprised to see ME Bank’s Microsoft-related costs rise as it grows its operations.

To be honest, I had thought that the days of Microsoft using selective material from favourable case studies to generate one-sided attacks on the Linux world were over by now. I mean, frankly, Windows Server is a huge property now; it and Linux have singlehandedly wiped out the traditional Unix and Novell server worlds, and from what I have been able to see, Windows Server has been making in-roads into the Linux server market as well. I don’t see many organisations at all making a big song and dance out of deploying Linux any more, whereas I see plenty deploying Windows Server, and increasingly, Hyper-V.

Issuing this kind of statement kind of makes Microsoft look a little silly. It draws attention to the fact that Linux is still a viable competitor to Windows Server, and implies that Microsoft is afraid of that threat. The reality is that in enterprise IT circles, the Microsoft stack is already dominant and getting more dominant every day, and I doubt that Redmond has much to fear at this point from the barbarian Linux hordes. Feel free to provide evidence in the comments below that I’m flat-out wrong; it’s been a while since we’ve had a decent Linux versus Windows debate on Delimiter.

Image credit: Ian Burt, Creative Commons


  1. Solaris isn’t linux, Renai. It’s a commercial unix OS. Commercial, meaning a paid product.

    If they looked at IBM DB2, or Oracle, clearly they’re going to pay for that. Both would likely only be officially supported on a commercial Linux Distribution, that also would carry costs.

    Without knowing what the actual alternative solution was, it’s scale and components, it’s a bit obviously one sided.

    Remember, if this is a financial business, they won’t be installing Ubuntu, MySQL and crossing their fingers. :)

    So there are going to be commercial costs, for a commercially supported linux based OS, running a commercial DB back end. That will always carry a cost.

    • I think people have forgotten that Linux was originally based on Minux, which itself was originally based on Unix.

      I guess they all have a similar command prompt though.

      • Linux was never *based* on anything. It was originally a Minix clone, but was written from scratch (because of the licensing restrictions of using Minix’s code). Torvolds used the GNU userspace with his kernel because licensing issues meant he also couldn’t use the UNIX userspace.

        This is unlike, say, BSD, which is a descendent of UNIX and contains UNIX code.

        • Being a bit pedantic don’t you think?

          From: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)
          Newsgroups: comp.os.minix
          Subject: What would you like to see most in minix?
          Summary: small poll for my new operating system
          Date: 25 Aug 91 20:57:08 GMT
          Organization: University of Helsinki

          Hello everybody out there using minix –

          I’m doing a (free) operating system (just a hobby, won’t be big and
          professional like gnu) for 386(486) AT clones. This has been brewing
          since april, and is starting to get ready. I’d like any feedback on
          things people like/dislike in minix, as my OS resembles it somewhat
          (same physical layout of the file-system (due to practical reasons)
          among other things).

          I’ve currently ported bash(1.08) and gcc(1.40), and things seem to work.
          This implies that I’ll get something practical within a few months, and
          I’d like to know what features most people would want. Any suggestions
          are welcome, but I won’t promise I’ll implement them

          Linus (

          PS. Yes – it’s free of any minix code, and it has a multi-threaded fs.
          It is NOT protable (uses 386 task switching etc), and it probably never
          will support anything other than AT-harddisks, as that’s all I have.

          • Is it just me, or does that Torvolds quote just say exactly what I said – Linux was an OS that resembled Minix but was free from any Minix code.

            When a programmer says something is “based on X”, they generally mean that it’s built on X’s code. Linux, being free of Minix code, was not based on Minix (or UNIX – which was completely separate from Minix again).

            You might say “based on” in that it was “based on some of the concepts” Minix was, but that’s not even really broadly correct – the OS resembled Minix but under the hood is quite different – Minix being a microkernel and Linux being monolithic/hybrid.

          • I personally thought his point was valid. Misconceptions or errors that go uncorrected is the reason we have to put up with Americans saying ‘aluminum’.

          • I hate using wikipedia as a source, but …

            MINIX is a Unix-like computer operating system based on a microkernel architecture created by Andrew S. Tanenbaum for educational purposes; MINIX also inspired the creation of the Linux kernel.

            When Linus Torvalds first started writing his Linux operating system kernel (1991), he was working on a machine running MINIX, so the initial releases based a lot of functionality on Minix subsystems.[2] Until the April 1992 introduction of the Extended file system, Linux used the MINIX file system.[3] The format is still used by some Linux distributions for bootable disks and other situations where a simple and compact file system is needed.

            So Minix inspired Linux, and the original versions of Linux used the Minux file system (the source for that statement is a book by Torvolds too by the way), and yet when I say Linux was based on Minix, I’m wrong? Seriously?

      • Solaris at least used to be owned by Sun for their Unix systems, last time I had any real involvement with it was about a decade ago so now idea who owns it now.

        • Solaris is now an Oracle product; it’s based on SunOS and runs on SPARC, as well as intel.

          Linux could be considered a distant cousin. Many Linux distro’s are “free” but if you want to run any in an enterprise, you’re going to turn to something with paid support.

          Doubly so if you want to run Oracle, DB2 or any other enterprise grade DB solution.

          Again, this is likely Microsoft tilting at Oracle. A long standing tradition, there. ;)

    • Solaris isn’t linux, Renai. It’s a commercial unix OS. Commercial, meaning a paid product.

      Oracle will happily sell you Solaris, and/or Linux, and/or MySQL and/or Oracle-Database. All of those are commercial, meaning a paid product. All are sold by Oracle Corp.

      If you are a bank, especially a small bank that can’t carry the support burden internally, you are going to be buying a support contract regardless of the product.

  2. Renai,

    1. $100k *increase* in license fee more likely to be from DB than Linux (RedHat or SuSE/Novell).

    2. Oracle is still increasing its revenues. Big in the Enterprise.
    Typically, I believe, on “mid” platforms between Mainframe & WIntel servers – ie Unix & Linux, though on all platforms.

    3. If you can lineup an interview, ask the Oracle view. They lost a Solaris and prob. DB account.


    Share Price movement since Feb-2000:
    NASDAQ -14% (fallen)
    ORCL +38%
    IBM +82%
    HP +62%
    Apple +1650%
    MSFT -33% (fallen)

  3. .. or Microsoft went in with loss-leading pricing on this run, just to ensure vendor lock by looking initially cheaper.

    OEL, RHEL and the DB2/ODB products, while not free, are not that expensive on the Linux platform. As for skills, I think Microsoft just threw them some free consulting, I’m sure the talent would have been there for the deployment if they looked rather being told by Microsoft.

  4. “joint statement with Microsoft” oh yeah this couldn’t be slanted at all now could it!

  5. (disclosure – I work for Red Hat, but about as far from sales and marketing as you can get)

    I suspect the key is actually earlier in the article: “historically it had depended on two core banking systems — a Windows–SQL-based product called NTBS that managed all savings and deposits; and a Solaris-based product called ULTRACS for loans and transactions products.”

    So, if they were looking to migrate to a consolidated system, it seems plausible that migrating *just* the Solaris systems to Windows could be cheaper than migrating from a mixture of Solaris and Windows to purely Linux.

    And then once you look at, it seems likely that the competing Linux solution was Oracle-on-Linux. If the real story is “Running Microsoft’s SQL Server is cheaper than running Oracle’s database”, that would surprise exactly no-one…

    • Unless of course you were looking at total cost of ownership against capability offered – there’s a reason there are so many Unix/Linix Oracle shops out there in the corporate finance world. It’s because it’s a serious , proven and reliable solution that scales.

  6. I find the extra cost claim very vague. Extra $100k over what period? Comparing what, exactly, with the selected solution? How were the extra costs divided between licenses, expert technical implementation and support fees, additional delivery overhead, what?

    Look, I completely agree that Linux experts are far rarer and cost top dollar compared with the far more common MS certified professionals, but that doesn’t mean you can’t find an IT company to deliver and support a Linux environment for a similar annual cost to a Microsoft alternative (particularly if you’re making use of GUIless Windows Server, which tends to confuse a lot of lower pay grade techs, but is the best ‘apples to apples’ alternative to Linux).

  7. There’s a reason that Linux skills are more expensive that Windows. Linux admins have to know what they’re doing. You can’t BS your way through Linux administration.

    For many Microsoft MSCEs that I’ve met over the years, this is the sum of their problem-solving skills:

    1. Restart the application
    2. Restart the OS/reboot the box.
    3. Deinstall/reinstall the application
    4. Deinstall/reinstall the OS and all of its applications
    5. Log a call with Microsoft support and wait.

    • Good Linux person would be claiming top $, so is MS person.
      also lack of wide Linux distribution means lower number of person on the market, that leads to higher costs of hiring.

      Its NOT because of lack of knowledge on either platform.

  8. For many Microsoft MSCEs…

    You have a tyop there.

    It’s MCSE – Minesweeper Consultant and Solitaire Expert.

    No no no. That’s not funny.

    It stands for Must Consult Someone Else……

  9. $100k over 5 years?

    Even for a small bank that is a rounding error in the IT budget.

    Not saying that it isn’t worth having, but it wouldn’t have taken much for the other option to come in cheaper, especially when you consider internal staff costs are probably budgeted at $100+ per hour (including all employment costs)

  10. Wow it’s been a while since I have seen one of these fine examples of

    hmm “Marketing”

  11. Why would you even mention MySQL in the article? PostgresSQL is where the comparison should have been if we’re talking about competitors to MS SQL on the open source side of things.


  12. Oh dear. If anyone tells you they can cost a major software platform over five years to $100k they are lying to you and possibly themselves. This is 20% of a support engineer or part of the upgrade cost when the thing starts to clag due to added functionality or unanticipated requirements. No one would actually prove this kind of crazy claim – they just put some estimated component costs in a spreadsheet and kowtow before it. Five years down the track, a whole bunch of scope changes have been made that render the original calculations moot. You can bet all your limbs that Microsoft and BE Bank won’t be releasing a nice joint presser in five years saying that their cost estimates were correct. A “good decision” was made and they’ve moved on.

  13. “Although one operating system option was essentially free to deploy, based on our past experience, we knew that it would cost more to support than Windows.”

    Since they use Windows/Solaris in-house I was wondering where this apparent past experience with Linux comes in?

  14. as mentioned above, most MCSE’s use google as their primary source of troubleshooting information, and cross their fingers that something serious doesn’t happen. however they are widely available, compared to experienced Linux admins, and retraining ME Banks’ existing support team may account for the relatively trivial amount of $100K.
    also, bear in mind the shortage of skilled IT staff which is pushing more outsourced labour and 457 visas into our corporate sector — this may influence the decision rather than rough cost estimates — local admins with Linux skills are going to become increasingly difficult to come by, so Windows is probably a safer option, albeit vastly more expensive.

Comments are closed.