On thanksgiving in the US, there is much to reflect upon. Introspection: what you are doing right, and what you are not, is always good. We do quite a bit of it. Though on Thanksgiving, it is interspersed between the mashed potatoes, turkey, and other elements.
HPCwire appeared to do some introspection. Sort of. Their language and adoption of one side of a debate is, well, troubling.
They cherry picked from John Power’s blog. He did in fact write a fairly good and thoughtful piece. Albeit one with an unfortunate title “SC07: the basher’s ball”.
Understand that John and his company sell a product that depends critically upon people being willing to adopt Microsoft products for HPC work. This is not a conflict of interest, John is not representing himself differently, and his product is gaining traction in specific market sub-verticals.
I might venture that it is not gaining the traction he thought it would though. This is a guess, and it could be wrong. Could be badly wrong.
There has been a perception from some quarters that once Microsoft entered the market that we would see a mass migration over to their platform. And products like John’s would take off in a wildly successful flight.
One thing about product marketeers and planners … they like hockey sticks, and they drink lots of koolaid. I still get a kick out of the old Gartner graph showing adoption of Itanium platforms. This was amusing at best, pointing to extreme failures to understand the real market on the part of people selling, building, and working on product. Hockey sticks galore.
Itanium is actually a fairly good model for this discussion, it’s trajectory, while not an abject failure, would be very hard to classify as a “successful” even if you had the luxury of re-defining what “successful” meant.
But this was a side note, back to the things which bugged me.
John points out in his blog that lots of people have a negative view of Microsoft.
I agree with this. There are many who will get green around the gills when Microsoft is mentioned. As I have noted many times, there are Microsoft tools that you would have to pry from my cold dead fingers (Excel and Powerpoint). OpenOffice is good, but if Microsoft released a Linux version of Office, I would buy it.
Big, massive hint to Microsoft on that one. There has been other talk this past week on the large and growing (and for the most part ignored) linux desktop market. Looks like it is larger and harder to ignore than most people thought. This is a market in which Microsoft could make money, by releasing product to address. Powerpoint and Excel. I can run them now via other modalities, but I would like them native. 64 bit. Please?
FWIW: Microsoft’s competitors in this space are releasing desktop products for Linux this way. If there wasn’t a market, if there weren’t demand, they wouldn’t do it.
Back to John’s points.
I understand it. Its a combination of dictating solutions to the market, bullying tactics, and marketing agreements that preclude competitors, to the point where you must pay what amounts to a tax on every machine you buy, whether or not you plan to use their product on it … Kudos to Microsoft for negotiating these. Great job. Just remember that when you are looking to buy some machines, and you want control over what OS runs on the machine … and you cannot decouple the purchase of the machine from the purchase of another license of a particular OS … which might not be the OS you had wanted.
Some people don’t like that.
I do agree though; some of the software is absolute garbage, and some is quite good. I just wish that the corporate ego wasn’t so tied into their OS platform that it precludes them from offering it under other systems. MacOSX which is a Unix-like system (very similar to Linux in many ways) has a fully operational Office. Why not Linux? The reasoning isn’t technical. Its marketing. Pure and simple. Corporate ego.
Some people I know are ardent bashers of Microsoft, some are ardent supporters. We take a pragmatic business focused approach. Where Microsoft products make sense, we advise in this direction. Where they do not, we advise in other directions. We are highly concerned with emergent issues in security, performance, and interoperability. These are not areas that Microsoft has not done a good job in. We are unimpressed with their marketing FUD and sabre rattling. If they have patents that cover technology being shipped, they need to identify it, which they have been unwilling to do to date.
Microsoft does make products. Many products. That span wide industries. They are trying to out iPod Apple. They are trying to out Symbian/BlackBerry Nokia and RIM. They are trying to make servers, and desktop systems. They want to out google google. And out … (and so on).
Our customers want Microsoft products to play well with non-Microsoft products. They want interoperability. They want security. They want performance. They don’t want added cost, complexity, and so on.
We as a company want to lower barriers. This has become our mantra as of late. We see too many barriers being erected to use of HPC technologies, and we want to see them reduced or eliminated. This is part of the reason we are developing DragonFly. This is part of the reason DragonFly is OS agnostic. This is part of the reason that we are interested in ways to lower barriers to people choosing whatever OS they need on an application by application and run by run basis. The OS is a detail, not a decision, though due to costs, some of them are decisions.
So far, John and I aren’t too far apart. I disagree with some of his language, or put another way, we see lots of Linux bashers running about, talking in the press. We see labeling of Linux as a cancer, by Microsoft. We see the sabre rattling.
Two sides to the same coin.
John concludes with
I would venture to say I would like to hear what “greatly” is defined as in this context. I am not sure we have the same definition (remember, I am pragmatic, if Windows was taking the world by storm in HPC, we would be backing it whole heartedly … same with Solaris, *BSD …. whatever).
That nit-picked, I agree. The HPC initiative should be evaluated in a larger context. In the context of the market, of demand, of value, of everything customers care about. The competitive solutions should be addressed similarly.
Johns final sentences are interesting
Allow me to note that Linux bashing, in active gear on the part of Microsoft and others working with Microsoft, does nothing to elevate discourse, and badly reduces it. This doesn’t excuse bashing your competitor. But when the CEO of the company producing the product is calling his competitor a cancer, publicly FUDs about IP infringement without indicating what is allegedly infringing, well, its hard say that “we are being bashed” without noting that you are in fact bashing. I am not talking about John in particular, but about the whole (quite insulting) initial marketing campaign from Microsoft. They did a really poor job of winning over hearts and minds.
Well, it of course, gets worse.
John’s article was reasonable. I didn’t agree with everything, but it was a good read.
The problem is that the “media” and various “analysts” have been using particular code words for Linux and Linux users for a long time. Unwelcome code words. Words that lower the discourse.
Rob Enderle was one of the major creators of this meme. Paint em as zealots, as they have a religion they are unwilling to compromise on. Read the linked article, you will see what I mean.
Sadly, Laura Dido and others in the media/analysis community do similar things. We expect very little in the way of objectivity from them, and our expectations of their output are met fairly regularly. They are the most vigilant and vocal in the Linux-basher crowd. They and their fellow travelers like to paint Linux and Linux users with broad brush strokes.
All of this was setup for the next portion.
Todays' HPCwire has an article on HPC Ideologues.
As background, Linux clusters have driven and continue to drive double digit year over year growth in HPC. HPC is now a majority in Linux clusters. They have been driving the business since 2003 (and arguably before that). They are responsible for driving the tremendous growth in this market, and the growth of HPC in general. It is hard to find fault in IDC’s numbers in this regard, they are right on the money. HPC has been transformed from a hodgepodge of proprietary technologies into an open set of Linux based systems. Linux clusters are the market. This isn’t boasting or bragging, go read the IDC data.
So the end users that procure this stuff are for the most part, Linux users.
HPCWire notes this itself
Yes. It has been. And shows no signs of letting up. It is creating new markets. And destroying old ones.
Unfortunately, the article goes down a slippery slope. Starts repeating marketing quotes and sound bites from Microsoft. In the next paragraph
On the contrary, it is as free as you wish it to be, and more importantly, you have control, not Redhat, not Novell, not IBM, not …
You can buy no licenses and deploy it everywhere (free as in “free beer”), and you can make modifications and redistribute them (free as in “freedom”). I cannot make a customized version of Windows and deploy it. It would dilute the windows brand.
But that mis-interpretation of “free” is fairly common in competitive literature, we run into this all the time.
We don’t expect to run into it in an article in HPCwire.
Moreover, the next paragraph, roughly a sound bite press release for Microsoft starts out with
And goes on from there. Apparently no one seems to have told the good people at HPCwire that you can in fact get all of what they are talking about today (and have been able to for a while) in a non-Microsoft platform. That is, they adopted the marketing language of one side of this “discourse”.
Of course, it continues
So, what does this imply? That they will eventually confront the Linux crowd?
It’s not rocket science to look at the words and see the implied future threats. Its not paranoia. Its marketing.
The real battle has been occurring for years, and by all objective measures that I am aware of, Linux is actively making headway and increasing market share on all fronts. Not just HPC. Some months/quarters are better/worse than others. This is a market, and the battle has been joined.
The article concludes with
Yeah, looks like HPCwire will need to rename itself soon. They have taken that, seemingly, all Linux users must be ideologues. I didn’t see them say that, but the implications are that if you are critical of the product offering (we are), that you must be an ideologue. And therefore a fundamentalist.
I think the original version I read said “zealot”.
Well, we see how HPCwire has aligned itself. It is firmly in the camp of one of its major advertisers. No surprises here.
How about this. Lets drop the Linux bashing. Lets drop the pseudo-IP claims. Lets work on interoperability. Lets work on lowering barriers. Lets work on making it easier for customers to do work.
Articles like the marketing material recycling one we linked to don’t help. Open minds do. We are looking at how to lower the barriers to customers using windows clusters. There are barriers, and they are not insignificant. Microsoft still has lots to do in market definition, pricing, and other issues. But we want to give our customers a choice, to lower barriers.
HPCwire did themselves, and their very large Linux using readership a disservice with this article. I hope they do reconsider their approach moving forward: as with the NYT claiming it is a paper of record for news, it is often derided as being overtly biased in a particular direction. Lets hope HPCwire doesn’t slip down this slope.
For the record, I remain skeptical on the business model for the windows clusters, but I think I see how to make it work (with some support from Microsoft). I remain skeptical about seeing large numbers of them ship (but I think I see how to enable them to be created as customers need).
The Microsoft people I met are passionate and focused. I like them. I am skeptical about their products success without some assistance, and I think I know how to provide it. But everyone needs to stand down from the attack postures, and focus upon providing value, and competing on ideas and capabilities. We need to lower barriers for our customers. Not raise them.
Our DragonFly product will be available on/for Windows, and it likely would work well with the Digipede product. It will work on Linux. It will seamlessly allow people to hop back and forth. The OS is a detail for a job. Not a religion.
n.b. I am as critical of Linux and Linux companies as I am of Microsoft and windows. I point out many broken things in Linux, and grumble about them, and how to fix them.