Saturday, February 13, 2010

An Answer to David Linthicum’s Questions about the Prospects for Windows Azure in his 2/12/2010 Podcast

David Linthicum is an early thought leader in the cloud-computing space, the writer of InfoWorld’s Cloud Computing blog, the author of Cloud Computing and SOA Convergence in Your Enterprise: A Step by Step Guide, and the newly appointed Chief Technical Officer of the Bick Group who acquired Dave’s Blue Mountain Labs.

Here’s his bio from the Leveraging Information and Intelligence blog on ebizQ:

David Linthicum is the CTO of Bick Group, and an internationally known distributed computing and application integration expert. He has twenty years of experience in the integration technology industry, most recently as chief technology officer (CTO) at Grand Central Communications. He has also served as the CTO at both Mercator Software and SAGA Software, and has held senior-level management positions at Electronic Data Systems, AT&T; Solutions, Mobil Oil, and Ernst & Young LLP. He has consulted for hundreds of major corporations engaged in systems analysis, design, and development, with a concentration in complex distributed systems.

Dave describes his 00:15:24 Can Microsoft Make it in the Cloud? podcast of 2/12/2010:

Dave talks about Microsoft's recent release of Azure into production, and their ability to make it in the clouds.

Some assertions he made about Windows Azure in this post surprised me, so I had the MP3 file transcribed and have added my comments about his conclusions here.

Windows Azure and Office Online?

Dave concentrated on Microsoft’s forthcoming Office Live Workspace for much of the first six minutes of the podcast

01:40 I’m not sure if anybody noticed, but you know Microsoft Azure, which as been been in beta for a while, in other words, people have been using it, they just weren’t billed for it, went live this week. So in other words, they turned on the billing. It went into billable production last week after a pretty short beta period by Microsoft standards. …

02:26 But what was interesting to me is that, you know, number one, they're kind of late to the party in terms of cloud computing, which is a typical Microsoft kind of thing to do. And ultimately, I don't think that a lot of people in the world of cloud computing really gave it much of a thought. And, you know, I predicted, you know, back in December that I think Microsoft will have a bigger impact on the cloud computing market this year and big splash mostly around the "Office in the cloud" products that we'll talk about later in the podcast.

02:51 But in was interesting to me how lack ... how much Azure really just didn't set the world on fire. I thought there would be more fanfare around Azure. So, ultimately, Microsoft's late entry into the marketplace kind of creates the question whether Microsoft will be able to catch up with the other infrastructure and cloud platform offerings or will its success in the cloud be limited to the forthcoming cloud versions of its popular office products. …

An apparent lack of fanfare for Azure’s release as a commercial, billable product was Dave’s theme in his Microsoft Azure is available, but does anyone care? post of 2/9/2010 to the InfoWorld Cloud Computing blog. In fact, the preceding podcast excerpt is almost identical to the InfoWorld post’s text.

I doubt if Microsoft will attempt to gain notoriety for Windows Azure among IT professionals and .NET architects/developers by promoting it as the infrastructure for Office Live Workspace. Microsoft is gradually moving much of its Web-accessible content to the same Data Centers that Azure uses, but I doubt that Microsoft or potential Azure customers would equate widespread internal use with success of the Windows Azure platform. Microsoft has directed its Azure promotion activities at .NET/Visual Studio developers (at PDC 2009, MIX 08 and 09, and similar conferences) and the Microsoft Partner Network (e.g., Azure Partner Hub). Microsoft has made and enormous investment in IT and developer evangelism for Windows Azure and SQL Azure.

Microsoft on the Desktop but not in the Data Center? 

The italicized statement below is preposterous in the light of the technical and business success of Microsoft’s Server and Tools Division in the past 10 years:

06:06 So, the big thing is that Microsoft has been chasing the you know infrastructure for years and has had very limited success. You know, ultimately, people love to have Microsoft on their desktops but they don't really want it in the data center [Emphasis added]. So I suspect that very similar things are going to occur around the Microsoft cloud offering going forward.

So, Microsoft has a tendency to be very good at providing commodity products such as operating systems and databases and office automation systems and things like that. But when you get into real enterprise-y things where you have to have transactionality and high security and very special needs, and you have to have openness and different interface-compatible systems, Microsoft has a tendency to kind of fall by the wayside. And I suspect that that's going to continue to be the case, even within their cloud offering. So we'll see.

Ultimately it's up to Microsoft to figure out how to market their cloud offering within the space, and the fact of the matter is that they have access to all of these existing Microsoft customers and Microsoft shops, you know, quote-unquote, that are definitely going to check out Azure before they check out the other cloud offerings out there, because their religious beliefs are around Microsoft, not necessarily Amazon or another of the cloud computing offerings out there.

It appears that more people want Windows in their data centers than alternative operating systems for x86 boxes. Peter Galli writes in his Linux Losing Market Share to Windows Server article of 10/25/2007 for eWeek:

In 2000, Windows comprised about half of the server operating system market, followed by Unix and Netware at about 17 percent each and Linux reaching towards 10 percent, she said, noting that today Windows owns about 70 percent, Linux about 20 percent, with Unix below 10 percent and Netware barely registering.

Following is a table of percentage of total quarterly server revenue for Windows, Unix and Linux servers from the last four IDC Worldwide Quarterly Server Tracker reports:

Operating System

2008Q4 2009Q1 2009Q2 2009Q3
Windows 35.3% 28.9% 38.1% 43.0%
Unix 36.2% 33.1% 41.4% 26.9%
Linux 13.6% 13.8% 13.8% 14.8%

 The Azure Giveaway to the National Science Foundation

Dave contends:

07:26 So one of the ways Microsoft can catch up in the marketplace is to give it away. And in a story that was out this week, Microsoft basically decided to give away their Azure platform to the National Science Foundation. This is a quote directly ... let's see ... directly from the release: "Researchers will have access directly to Windows Azure for a period of three years along with a support team help researchers quickly integrate cloud technology into their research. ... So Microsoft is trying to catch up by providing some very much, I guess, pro bono, are giving away their product to folks with the National Science Foundation, grant recipients, to get them up and running and working and working their grants.

This is a great move because, in my opinion, because, number one those folks are very distributed. I've worked in that world for a while and their ability to use competing resources is going to be then running down to the corner Staples and buying disk drives and computers and things like that, and this provides them a common infrastructure, in this case the Microsoft Azure platform wherein they can do their research. The other thing, too, it's going to provide Microsoft very good PR in that the National Science Foundation does some great work and the grant recipients are doing some great work, and Microsoft can ride along with that and provide them with a computing infrastructure to make them successful, the more popular and the better they're going to appear to the public. So, you know, ultimately, as we mentioned in the previous... earlier in this podcast, Microsoft is trying to reinvent itself out into the cloud and is having some limited success in having people paying attention to them right now.

As I noted in my Windows Azure and Cloud Computing Posts for 2/12/2010+ post:

Dave [Linthicum] also asks on 2/10/2010 Can Microsoft Catch Up by Giving Away Azure? in InformationWeek’s Intelligent Enterprise blog:

“Microsoft Azure finally went into production amidst much fanfare. The long-awaited Azure platform is Microsoft's big bet in the cloud computing space, but can they catch up to AWS,, Google, and other existing cloud players?

“That said, it is no coincidence that Microsoft is basically giving away Azure to deserving government scientists for three years in a program between Microsoft Corp. and the National Science Foundation. …

“Once again Microsoft is late to the party. However, they continue to hold a special space in the hearts of many enterprises, a brand loyalty that most cloud computing providers just don't have. The concept here is to get as many users on the platform as possible, in the shortest amount of time. However, is that a good strategy for Microsoft?

“If you look at the history of Microsoft they seem to get into games late, and still win. Their entrance into the emerging Web in the '90s was almost kicking and screaming after the Microsoft Network was released. However, once they set their sites on the Web, they owned the browser market after only a year.

“The cloud is a bit different. Cloud computing providers have already established their presence in the market. It's going to be difficult to attack users who are already loyal to one or two of the larger players, that is... unless you're willing to give it away for free. …”

I wouldn’t call ~$100/month for a single small Windows Azure instance to be “Giving Away Azure.”

Early in the preceding post, Dave mentions “much fanfare” surrounding Azure’s move into production status.

My Funding Details for the Third (Azure) Cloud-Computing Grant to the National Science Foundation (NSF) post of 2/7/2010 notes:

Microsoft’s cloud-computing grant, which has achieved widespread publicity, follows these earlier cloud-services gifts to the NSF by competitors: 

  1. “A set of cloud-based software services supported by Google and IBM” (National Science Foundation Awards Millions to Fourteen Universities for Cloud Computing Research, 4/23/2009)
  2. “Access to another cluster supported by HP, Intel, and Yahoo housed at the University of Illinois at Urbana-Champaign.” (HP, Intel and Yahoo! Create Global Cloud Computing Research Test Bed, 7/29/2009)

It appears to me that Google, IBM, HP, Intel and Yahoo started the cloud computing give-away trend, at least to the NSF.

Microsoft Then and Now: Internet Explorer and Windows Azure Aren’t Analogous

What went on in the 90s isn’t indicative of what’s going on in the 21st century first decade:

So if you look at the history of Microsoft, they seem to get into the game late as I mentioned earlier and still win. We'll see if that's the case here. Their entrance into the emerging Web in the 90s was almost kicking and screaming after Microsoft Network was released. Remember back the day, they released the Microsoft Network to compete with AOL and at the same time the Internet took off, so proprietary closed networks, such as AOL and MSN became kind of entries to the Internet at that point. They were kind of commoditized very quickly. But shortly after that Microsoft owned the browser market after only a year.

So, I view the cloud a bit different. People always talk about penetration of Internet Explorer as a good way to gauge, or a good way to justify that Microsoft is ultimately going to be successful in the cloud. And I kind of push back on that a bit. So the reality of cloud computing is that subscription cost to the platform has very little bearing on the return on investment of the platform. Azure, like any other cloud provider, will have to prove productive in order to be truly cost-effective. It also means being open, you know, something that Microsoft has had issues with in the past and definitely has issues with with Azure, which is very much a dot-Net-ty kind of product and does not look like, you know, ultimately the release of Azure [that] Microsoft really hasn't changed its spots in terms of providing more proprietary than open environments.

I haven’t heard anyone other than Dave “talk about penetration of Internet Explorer as a good way to gauge, or a good way to justify that Microsoft is ultimately going to be successful in the cloud.”

Microsoft offers a Platform as a Service, so it’s not surprising they chose their own .NET Framework and Visual Studio development tool for Windows Azure. However, the Windows Azure team sponsored PHP, IronPython, and IronRuby as well as the Eclipse toolset as alternative development platforms/programming languages for Windows Azure. Windows Azure storage uses the open, RESTful AtomPub protocol for data queries, insertions, updates, and deletions.

Leveraging the Microsoft Customer Base and Dealing with a Proprietary Culture

Microsoft has cemented its developer mindshare with .NET, Visual Studio, and its recently added support for AtomPub, PHP, Python and Ruby in Windows Azure projects.

11:09 You know, that said, I think Microsoft will gain a large portion of their existing market share in the cloud from those who are Microsoft shops, as I mentioned earlier, are waiting for Microsoft to tell them how to go to cloud, so they can consume everything Microsoft produces and Azure is going to be nothing different. So that's actually going to be a good move, considering that Microsoft is bound to maintain compatibility with existing systems around .NET and the fact is that Microsoft can build some pretty impressive stuff, as I mentioned earlier, when they put their mind to it. So if you an existing Microsoft shop, going forward, Azure is typically going to be you best path to the cloud, or at least something you need to look at ultimately, because the .NET and backward compatibility and, you know, the kind of the very cool integration that Microsoft has a tendency to do with its own products will perhaps carry forward with Azure, we hope. And ultimately that may be your best path to cloud computing; at least something you should consider.

12:06 However, Microsoft, as I mentioned just a minute ago, has a tendency to do everything in a very proprietary way. So the ability to provide open systems, excuse me, open interfaces, and interfaces with other cloud providers, and interfaces and code compatibility with other on-premise systems, other than their own are going to be a challenging thing for Microsoft; very much a cultural change that they need to go with going forward.

It remains to be seen whether any cloud implemented as a PaaS will be interoperable across multiple operating systems and development platforms without at least some modifications to applications.

Microsoft’s Need for Innovation

Dick Brass’s Microsoft’s Creative Destruction article of 2/4/2010 for the NY Times evoked a whirlwind of op-ed commentary and an epidemic of navel-gazing by industry pundits, including Dave:

Microsoft is a changing company, so you go back in time 15 years when I was working at PC Magazine and working with Microsoft on doing OS evaluation and evaluating application servers and, you know the old MTS system and things like that – very, very much innovating in the space and leading the market. And ultimately, Microsoft has evolved over time to being less innovative, in my opinion, and, now that we're moving into the cloud computing era, ultimately innovation becomes a paramount ... innovation becomes a characteristic that absolutely needs to be present within Microsoft in order for them to win the cloud-computing space.

With the initial release of Azure, which I'm kind of lukewarm on, I think they have the potential of being successful. So they definitely have an existing marketing presence, an existing very impressive customer list that most software organization would cut off their right arm to get, and now they have the ability to go out and penetrate that market. But they're doing so a bit late to the party. So cloud computing as a concept has kind of been around for a long period of time and definitely Microsoft has been a player in there since we've really started talking about cloud computing, but as far as a real productized offering in this space, such as Azure, well that really came out last week and we've been talking about and playing with cloud technology for rather a long time.

What’s missing from Dave’s podcast is why he’s “lukewarm” about Azure, other than it’s “a bit late to the party” and not sufficiently “open,” presumably meaning interchangeable with other PaaS implementations. Unlike military weapons makers, I don’t believe that cloud providers are obligated to make or use only interchangeable parts. Commoditization by standardization of PaaS clouds at this early stage in their lifecycle would be counterproductive to innovation.

Damnation by Faint Praise

So it will be interesting to see if Microsoft actually has an impact on the market. If they're able to become a legitimate player; if they're going to be a true contender to the Amazons and the Googles of the world. And it's kind of a "wait and see" right now. I think it's theirs to lose; they have a customer base and their market base; they've put a lot of investment in this space to get it up and running. I think it really comes down to the really innovative culture of the company and the ability to step up to the plate and become the leader here in innovation and not necessarily trying to follow and react to the market. I think that's a killer for Microsoft. They're not going to survive in the cloud computing space if they think that's going to be their strategy. That's my two cents.

For better or worse, Microsoft has an impact on any market they enter, whether it be for desktop or server operating systems, databases, game boxes, or for-hire cloud computing resources.

I’m surprised that Dave didn’t mention SQL Azure in his podcast. If SQL Azure can demonstrate (as promised) that it’s scalable with relational database sizes significantly greater than 10GB, it will prove that innovation remains alive and well in Redmond and other Microsoft facilities around the globe.

blog comments powered by Disqus