Reviewing a Microsoft anti-Linux case study

Reviewing a Microsoft anti-Linux case study

Summary: Microsoft simply hung an anti-Linux label on a very carefully worded story about a pair of committed Microsoft partners, HP and Accenture, getting together with Microsoft to sell rather simple technology to a willing customer.

SHARE:
TOPICS: Microsoft
56
As most people know Microsoft has an anti-Linux program called "Get the Facts" featuring case studies arguing the Windows case. When one of those, wearing the title: London Stock Exchange chooses windows over Linux for reliability, arrived in my email last week, I was sufficiently intrigued by the relative reliability claim to read the thing.

Here's the summary quotation attributed to the customer: LSE CIO David lester:

 

"No other exchange is undertaking such an ambitious technology refresh programme based on next-generation Microsoft technology. We've always provided a first-class service, but now we can claim to be the fastest in the world as well."

Take a careful look at the actual wording: "No other exchange is undertaking.." and, "now we can claim to be the fastest in the world." (Emphasis added.)

The Tandem system this replaced was first installed in 1995 and had earned its non-stop trade name with zero downtime over the last six operating years, but now belongs to HP and is therefore going away. In response LSE CIO David Lester developed a plan - one structured around a partnership with Microsoft:

 

Before choosing Microsoft technology, the London Stock Exchange reviewed several potential architectures to meet the requirements of Infolect® and the TRM design objectives. The Microsoft .NET Framework -an integral component of the Windows Server® 2003 operating system- was selected for a number of reasons, including developer efficiency, performance, and scalability. The Infolect® application, which went into production in September 2005, was implemented on a total of 120 HP ProLiant servers across multiple data centres. This configuration allows Infolect to process an average of 15 million real-time messages a day distributed to more than 107,000 trading screens in more than 100 countries.

 

120 HP Proliant servers sounds like a lot - but then so does 15 million if you're thinking in terms of personal dollars or weeds to pull in your garden. On the other hand, the need for 120 servers is less than obvious given that the volume amounts to only about 520 messages per second for an eight hour trading day and is easily within scope for a small Unix server like a four way Opteron or T2000 - something that shouldn't be a surprise given that the application formerly ran on nothing more than a somewhat upgraded mid nineties Tandem.

That 520 messages per second average volume estimate, by the way, matches what you get if you look closely at their claim to be the "fastest in the world."

Here's their headline summary:

 

London Stock Exchange Cuts Information Dissemination Time from 30 to 2 Milliseconds

That's fast - in fact, 2ms barely exceeds communications latency for a typical PC NIC.

But look carefully at the wording, especially as repeated in the excerpt below, and you'll see how this miracle is achieved:

 

Reliability is fundamental to the London Stock Exchange value proposition for the market and continues to give its senior managers peace of mind about system uptime. There are approximately 300 customers who connect directly to the live Infolect system to receive real-time market data directly from the London Stock Exchange. The data disseminated from Infolect is then displayed on more than 107,000 terminals in more than 100 countries.

Notice that they imply that their data goes to 107,000 screens worldwide, but actually say only that it goes to 300 or so customer owned devices on their local area network - and 520 messages per second means they take an average of just about 2ms to deliver.

In other words the performance claim amounts to saying they have a relatively slow LAN and those "107,000 screens in 100 more than countries" have absolutely nothing to do with it because those networks are run by LSE's 300 or so big customers - and it's very doubtful that any of them would have noticed any performance change at all.

At the end of which, of course, I was still wondering what Linux had to do with any of this - Microsoft's headline, after all, said that the LSE picked Windows over Linux for reliability.

The answer is that neither Linux nor Solaris nor any other Unix variant is mentioned in this report; Microsoft simply hung an anti-Linux label on a very carefully worded story about a pair of committed Microsoft partners, HP and Accenture, getting together with Microsoft to sell rather simple technology to a willing customer - whose employers, I think, should be seriously embarrassed.

 

Topic: Microsoft

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

56 comments
Log in or register to join the discussion
  • A case study of *nix (folks)

    *nix folks remained with the same 70's ideas while the world has progressed.

    Microsoft constantly embraces new technology and hence is at the forefront of it. The London Stock Exachange is just the begining. If anyone who has tecnical or programming knowledge of .NET, Java, C++, Windows, *nix, computer security will know how far ahead Microsoft is when compared to their competition. Its only a matter of time before NYSE, Nasdaq switch over to Microsoft technologies. I wouldnt be surprised if more stock exachanges and more mission critical stuff switch from *nix to Windows.

    Its no coincidence that the saying goes "Its dangerous to have a little knowledge than no knowledge at all". Murph why dont you go hit the book and start reading. Your ignorance is showing.

    PS - I work at a company that does mission critical software development of distributed systems. I definitly know more than a little bit of whats happening here.
    zzz1234567890
    • Some things don't change, youngster

      On 11/21/06 defconvegas spoke and said:

      > Microsoft constantly embraces new technology and
      > hence is at the forefront of it.

      Having been forced to handle some business this weekend on a Windoze box which refused to boot any sensible LiveCD, I find your comments laughable. 6 reboots in an hour of an XP system is not my idea of reliable and I strongly appreciate the Massachusetts ODF advocates' desire for a file format for archival purposes they are not constantly having to upgrade. Doing one thing and doing it very well is an idea which is much older than the seventies, and it remains a good one.

      This particular Windoze network may make sense for the limited network Murphy is talking about. For the time being. Whether it remains so will depend in part on a company which advocates constant updating and upgrading refraining from the release of anything which is incompatible with it. Many of us do not trust them for that.

      It's not as if M$ doesn't have its place. It does. That doesn't change that this particular report is inherently dishonest without being an outright lie, which is typical of them, they are sophisticated in their FUD, or that your criticism of *nix is utterly beside the point. It doesn't matter how true it is, if people who appreciate the Old Ways are dinosaurs, so be it. So are birds: dinosaurs have been around longer than mammals and they are not extinct so they must know something.
      jplatt39
      • laughable

        [i]"a Windoze box which refused to boot any sensible LiveCD, I find your comments laughable."[/i]

        What's laughable is your comparing a solitary Windows XP machine running on who knows what hardware and maintained by who knows who with redundant, mission critical systems run by professionals.
        toadlife
        • Excessive redundancy

          120 boxes is a pretty redundant, and I'm sure it requires plenty of redundant professionals to keep it running.
          Erik Engbrecht
          • Then perhaps some of them should be made redundant.

            Sorry. British humour. Couldn't resist.
            dave.leigh9
      • *nix criticism

        ... your criticism of *nix is utterly beside the
        point. It doesn't matter how true it is, if
        people who appreciate the Old Ways are dinosaurs,
        so be it. ...

        I could not agree more!

        I was a McSoft systems admin for six years and I
        have the multiple scars to prove it. Then I dove
        into Linux on a whim. That led to a job working
        on real servers running *nix, BSD and Linux.
        troy15
      • umm...ok

        Well if that's true I wouldn't run Oracle on Linux either -- I'd be running AS/400 still. I'm curious why this author doesn't go after Oracle for doing the EXACT SAME THING in their advertising. Or IBM maybe, who makes some pretty extraordinary claims about their Linux based solutions? This is typical software advertising and anyone who buys claims made by a company for their own purposes without some scrutiny doesn't need to be making decisions. I think ZDNet should be embarassed actually for employing a writer who is obviously a technology zealot. How is this person supposed to accurately report the facts when he has a blatant predisposition towards one technology or against another?
        gigabot71
        • What does it matter what you run SQL on?

          I'm not saying that technology doesn't change. I'm suggesting that there are rights and wrong. Of course a faulty computer has very little to do with what the London Stock Exchange can afford, but it does illustrate very well what has always been one of M$'s weak points: that as products like XP near the end of their cycle they provide very little concern for backward compatibility and that it is certainly possible to break the installation with upgrades.

          Technology does change. And we have machines more powerful than the AS/400s. What doesn't change is that the important question is are they better used than the AS/400s? Are they?

          Since I'm writing out of exasperation, as opposed to trying to convince anybody, I'll ask the obvious question: who would you expect ZDNet to have write a "Managing L'Unix" column other than someone with a love of Unix? I hope he does have a "Technology Zealot" T-Shirt.
          jplatt39
      • that is just so tired.

        windoze. M$. That is SO old. And it points directly to a lack of decent objective intelligence. It shows a bias and the world of IT is a profession, not a religion.
        I don't see people posting naive, small little plays on words about other systems here.
        xuniL_z
        • I agree

          Otherwise I'd be overusing

          EUNUCHS

          because it's just so appropriate

          ;-)
          TonyMcS
      • Very true. Try this experiment.

        Take two server boxes. One with half the CPU speed and RAM of the other. Install any server Linux on the 1/2 box. Install Windows 2003 server on the other.

        Use each of the two for 3 months.

        You will notice that the Linux box will still out perform the Windows box, not need rebooting, and will not have any security problems.

        Now try taking off the antivirus, spyware program and firewall off both.

        Now try them both as a desktop system with novice users.

        This is the only testing that really matters.
        slim-01
        • Only if you try this experiment

          Put down the kool-aid.

          Step out of your reality distortion field for a moment and take in the scenery.

          I recommend only short trips into reality at first, as the initial shock might be disturbing.
          toadlife
          • I have actually done this with a Linux box 1/4 the Windows box

            Same results.

            Have you haven't actually done this test then you comments are the reality distortion.

            Your comment sounds like "That doesn't sound true so let me make him sound like an idiot".

            I have also tested a Windows 2000 server compared to a Windows 2003 with twice the power and Ram and even though it wasn't as dramatic as the Linux comparison it was very noticeble. The lesser Windows 2000 box out performed the Windows 2003 box.

            For the record I was being kind in my first post. The truth is that I couldn't wait to get my Linux Server power back. The 4 year Windows test was dismal failure in comparison to my Linux server.

            How about Comparing Linux downtime and reboots to Windows downtime and reboots. 1 Linux to 200 Windows. Again these are my own actual results. Also the 1 Linux reboot is for power falures/new hardware.

            This is based of 4 years using a Linux server and 4 years using a Windows server. As far as I can tell this test is pretty much the standard results from people who have actually used both.

            This test also helps explain why so many Windows Tech people are afraid of Linux. Job Security.
            slim-01
    • If that is true why is every Accounting system programmed with FoxPro crapp

      if compared to other systems like Oricle based ones.

      Not comparison. FoxPro based packages are very weak and crash often in comparison.
      slim-01
      • Same reason people don't drive tanks to work.

        Sedans are very weak and crash often in comparison.

        How much are you going to pay for security and stability? Even when the stakes are your life, as with buying the sedan, you make that trade-off.
        dave.leigh9
        • In my experience non FoxPro accounting systems are not cheaper

          I have used two FoxPro based sytems, two Oricle based and 4 proprietary based.

          Unless a person makes the mistake of building a custom system and wind up doing the beta testing on the new system the last two are about the same price. They of course pay up to a couple million on software that never will run as well as off the shelf accounting packages.

          The only system that tend to be cheaper are small business software like QuickBooks which I like except for their lack of a Linux/BSD/Unix version.
          slim-01
          • Opps I meant FoxPro accounting systems are not cheaper

            Sorry about that.
            slim-01
          • They ARE inherently cheaper to build. That has little to do with price.

            I have experience as a FoxPro developer; specifically, one that has worked on accounting systems. The fact is, you don't need Oracle or any other back-end database system to create a FoxPro application. Nor do you need it to deploy. FoxPro scales from a single-user to multi-user system with exactly zero need for additional server licenses or server software. Foxpro systems are pretty damned cheap to build and deploy. What you're [i]charged[/i] is a whole different ball of wax.

            Don't make the mistake of thinking that I disagree with you about off-the-shelf packages. Accounting is a very mature market... one that hasn't had any real innovations as far back as you care to go. For instance, if you're running an SMB on an SBT Database Accounting package you bought 10 years ago (patched for Y2K) you have no need to upgrade unless you just want to. (And by "no need to upgrade" I mean far into the forseeable future, and that goes for any mature accounting system scaled to your needs.) The ease of incorporating vertical market software may have made the choice of a customizable system reasonable at the time of purchase; just as the cost of swinging those vertical packages to a new system may make a change prohibitive today. (Back when I was selling accounting systems, the choice of system hinged primarily on the need for vertical market integration.)

            The problem with the price of accounting systems in general does NOT stem from the development system on which it's built. Rather it is a sad reflection of the economics of corporate sales. Companies practically beg to be taken advantage of, and the prices in this market are scaled up accordingly.

            Partly this is due to the fact that corporations require service contracts that often far exceed their needs, and certainly exceed the ability of a company that caters to the SMB and above market to absorb through volume. The vendor has no choice but to set a price level that covers the customers' maximum demands. Those demands are rarely realized; so the vendor learns to depend on service contracts as core revenue. It's practically free money.

            Thus, even a cheaply built "enterprise" system is overpriced due to the need of the vendor to CYA wrt the service contracts. This contract overkill leads to a self-perpetuating system where the vendor charges a premium because that's what the market will bear, and the clients accept that kind of price level because that's what the market provides.

            (If all that sounds crazy, imagine trying to sell at a reasonable price in that market, only to find that your bid is rejected for being too low; then make the sale next time by arbitrarily doubling your price.)

            OTOH, shrink-wrapped packages simply need to ensure that they are reasonably free of bugs, and may distribute the same changes to all users on the vendors... [i]not the clients[/i]... timeline. That makes a world of difference in terms of price.
            dave.leigh9
          • My main probrem with FoxPro based Accouting software

            They crash and require reboot all the time. They are prone to recurring errors. Even on brand new top of the line boxes with plenty of Ram.

            If you call the company making the program and they will tell you don't run any other programs, don't let the antivirus check the exe of the accounting program, etc...

            I have used several FoxPro with similar results. I have run Oracle, MySql and proprietary accounting systems with 180 degree results. No lockups, run a dozen other programs and all of these had features coming from the database at the base level. Things like being able to lookup anywhere, edit the lookup, print the lookup.

            After years of assuming that it was the accounting software company I finally saw a trend and did some studying and found out that it was the base system that was the deciding factor on how good the final program was.

            Now if someone as me about accounting programs the first thing I ask is you need to find out if it is FoxPro based. If it is keep checking into other options.

            There is one option that is far worse and that is having a program designed for you company. I have never seen out of the several dozen companies try this have a good finished product.

            If you stop and think about off the self solutions have other people do their beta testing. When you have a custom built accounting system only one company is doing beta testing on that product.

            First I would never beta test an accounting program on live company books. Second one company isn't enough. Let someone else do the risk of beta testing.
            slim-01
          • Different experiences.

            I agree that if you have the option then you should go with an off-the-shelf solution scaled to your needs. Accounting is, after all, much the same the world over.

            However, you and I have had completely different experiences with Foxpro programs. 'Nuff said.
            dave.leigh9