September 2011

Windows 8 demos at Microsoft’s BUILD developer and partner conference have been very compelling, inspiring even. But nothing about the UI will change the underlying challenges with Microsoft’s open ecosystem. Users will still have to deal with frustrating experiences, even if the blue screen of death is replaced with a blue frowny face.
Windows 8 looks promising
Our own Galen Gruman’s review of Windows 8 is quite glowing. Galen goes out on a limb and suggests that HP’s decision to jettison WebOS could have been due to Windows 8:

But if Windows 8 is nearly as good as the demos look, Microsoft could very well win the mobile wars, despite years of failures in Windows tablets and mediocre smartphone efforts. If Hewlett-Packard CEO Léo Apotheker had seen a preview of Windows 8 tablets, that would explain why he suddenly killed the WebOS-based TouchPad tablet last month.

Other reviews of Windows 8 have been cautiously optimistic that Microsoft may finally have an OS to combat Apple.

The only problem, software alone is not enough. The real test for Microsoft is how Windows 8 will demo on the hundreds or thousands of devices, PC and mobile, that will be “optimized” to run Windows 8. I stress optimized, because every hardware vendor will play that card, when in fact, no piece of software can be optimized for everything. That’s where marketing and reality depart.

Configurability versus design choices
John Gruber wrote a thought provoking post about Apple’s long term sustainable advantage residing not solely on their design, but their supply chain. The two points are related, and will impact Microsoft’s windows 8 strategy, especially as they grow beyond the desktop to tablets and mobile devices with a single operating system.

Gruber wrote:

Design is largely about making choices. The PC hardware market has historically focused on three factors: low prices, tech specs, and configurability. Configurability is another way of saying that you, the buyer, get a bigger say in the design of your computer. (Bright points out, for example, that Lenovo gives you the option of choosing which Wi-Fi adaptor goes into your laptop.) Apple offers far fewer configurations. Thus MacBooks are, to most minds, subjectively better-designed — but objectively, they’re more designed. Apple makes more of the choices than do PC makers.

I’ve been thinking of this more and more as part of my day job, and I can fully understand why making choices are hard for vendors. Clients tell us that they want to make choices, because a lack of choice can sometimes lead to vendor lock-in. But these same clients demonstrate higher satisfaction with products which have been, in Gruber’s words, more designed, and hence present fewer choices to buyers.

Microsoft’s issue with Windows is that their OEM partners offer a degree of configurability that, on the surface is helpful, but turns out to hurt user satisfaction with both Windows and the hardware OEM.

I hadn’t made this connection until I started to use Windows 7 in a VMware Fusion virtual machine on a new MacBook Air. Yes, I know, the horror. But I need to use Windows for work and will be travelling with the need for my work and personal machine. This was easier than lugging around two physical machines.

Even with the overhead of a hypervisor and the relatively mediocre Intel core i5 CPU, my work hypervisor, is a delight to use. I’ve had no issues with driver mismatches or blue screens of death. Windows startups, shutdowns and resume from sleep are speedy, thanks to the SSD drives. I actually like using Windows again. More importantly, my PC is no longer getting in the way of my productivity.

For once, a hardware provider that’s actually enhancing satisfaction with Windows. Unfortunately, Apple isn’t a Microsoft hardware partner.

What’s Microsoft to do?
It’s difficult to know how Microsoft will address this issue going forward.

Microsoft could get very, very, restrictive about configurations and testing before allowing hardware OEMs to use Windows 8. This would require the same level of testing for fixes and upgrades to drivers used by the hardware configuration. However, considering the billion odd users of Microsoft Windows, with vastly different amounts to spend on PCs, a very restrictive policy will be at odds with Microsoft’s business goals.

Increased restrictions could encourage Windows OEMs to build with Linux OS, or more likely, Google’s Chrome OS. Microsoft is in a difficult spot of being the undisputed market share leader, but at risk of market share loss to Apple at the high end and Chrome and Linux at the low end. Until recently, the high end and low end competition was theoretical at best, but no longer.

It’ll be interesting to see what Microsoft and its partners will do if Apple uses its supply chain and lower configurability to offer a much lower price point entry to their desktops and laptops. In some respects, the iPad is doing just this as it eats into existing PC share.

Whether Windows 8 will be enough to stop the share loss is an open question. The real question however is how well Windows 8 will be configured and optimized for the hardware you’ll be asked to buy. Keep that in mind as you purchase new machines for your teams and employees.

I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies, or opinions.

HP announced intentions to take on Amazon in the public cloud infrastructure as a service (IaaS) arena. However, the beta announcement has little to no information about why anyone should consider HP Cloud over Amazon and other public cloud IaaS providers.

Little to differentiate HP Cloud Services thus far
HP’s recently launched beta of HP Cloud Services provides users access to two initial offerings, HP Cloud Compute and HP Cloud Object Storage. HP describes the beta as an opportunity to try these two services “through our easy to use, web-based UI on top of open, RESTful APIs, based on HP’s world-class hardware and software, and OpenStack technology.”

These two cloud offerings compete directly with Amazon’s Elastic Compute Cloud (EC2) and Simple Storage Service (S3) cloud services.

In describing HP Cloud Compute or HP Cloud Object Storage, HP makes no claims about why a company, ISV or developer should be interested in HP’s public cloud, over Amazon’s AWS cloud or other alternatives.

Enterprise-grade SLAs, management and monitoring, or hybrid cloud support, or differentiated pricing would all have been areas that HP could have used to differentiate HP Cloud Services.

But no. Instead, there is a seemingly random point about the HP Cloud being based on OpenStack technology. A point that received a lot of press, mind you. But let’s look at the reality here. HP joined the nascent OpenStack project on July 27, 2011. Knowing a thing or two about launching products within a large company, it’s very difficult to believe that HP could have altered their HP Cloud offerings in a meaningful fashion in a month.

HP’s cloud blog does make the OpenStack effort a little more real. HP’s Emil Sayegh writes: “HP developers are already active and many of our ideas will be shared at the upcoming OpenStack Design Summit and Conference, of which we are a sponsor.”

At this point, the OpenStack linkage with HP Cloud Services seems like a distraction. Hopefully this will change over time.

Why HP didn’t make more reference to its monitoring and management capabilities, areas where HP could clearly differentiate itself with Amazon AWS, is an open question. It could be that HP is targeting the broad market, and is less interested in enterprises at this time. The fact that HP is requiring a credit card for billing could be a tip here.

Asking clients interested in public clouds why they’re not using Amazon AWS today, I’ve often heard responses to the effect: “because my IT department doesn’t run on a credit card.”

Billing through a credit card absolutely lowers the bar to entry to HP Cloud Services. But it also turns of enterprise IT departments.

HP’s silence on pricing poses barrier to entry
Staying on the pricing thought, HP states the following on its website:

Stay tuned for information on pricing. We’ll communicate more before we begin charging for services.

Developers and enterprise IT should be concerned about devoting time to HP Cloud Services before pricing is known. It’s curious that HP decided to launch the beta without any pricing details just weeks after Google faced developer backlash after substantially raising prices once App Engine left preview mode.

Considering HP’s enterprise software and hardware heritage one could argue that HP will price higher than Amazon’s AWS, but offer higher value to enterprises. However, the focus on broad based developers, and requiring a credit card access to the beta, suggests aggressive pricing versus Amazon’s AWS. We’ll have to wait for additional information from HP to know for sure. If that makes you uncomfortable before approving proof of concept usage of HP Cloud Services, you should be.

Ask for clarity before devoting your time
Taking my vendor hat off for a minute, it’s absolutely within your rights as buyers and users to ask vendors, HP in this case, for clarity before making investment decisions. You and your teams have too much on your plates to work on proof of concepts without understanding how your business will benefit and what it’ll cost.

I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies, or opinions.

As with any new trend, hyped up by vendors and pundits, developers and CIOs interested in cloud must invest their time and budgets cautiously. Even with all the great new product and vision announcements at VMworld and DreamForce this week, two announcements will make it more difficult for developers and CIOs to leap into their next cloud investment with confidence; Google, VMware and, three vendors vying for cloud leadership status, share the blame.

Preview pricing has no place in the enterprise
Google products are well known for their beta status well into their public life cycles. The beta, or preview, moniker is fun and cutesy, until you’re trying to establish an enterprise foothold, which Google App Engine is trying to do.

The problem with betas and previews, aside from the lack of SLA support for enterprise production workloads, is the uncertain pricing associated with pre-GA products and offerings.

This point became crystal clear when Google announced new pricing for its App Engine cloud platform. The Hacker News and Google Groups message boards dedicated to App Engine are filled with developers complaining about dramatic, anywhere from 50 percent to over 2800 percent, cost increases. Speaking of the individual facing a 2800 percent cost increase, he writes: “we are moving 22 servers away. Already started the process to move to AWS“.

Amazon Web Services appears to be the beneficiary of Google’s new pricing announcement. Enterprise developer and CIO confidence in using pre-GA cloud services definitely take a hit with Google’s new pricing.

Complex cloud pricing poses a barrier for enterprises
It’s been said before that Google, for all its greatness, just doesn’t understand the enterprise software market; take a look at the current App Engine pricing model for proof.

Pricing per usage of bandwidth or compute instances is increasingly well understood by IT. In fact, these were the key elements of the original App Engine pricing model when the service was still in preview mode.

Pricing for five different API uses, as Google has introduced with the new App Engine pricing, is overly complex at best. Does the priced API model better reflect Google’s costs, and provide developers and CIOs an opportunity to reduce their costs by using cost effective APIs? Yes. But it’s also confusing and complex. In some respects, the new pricing model feels like Google let really smart engineers, or actuaries, set the pricing model as a fun math exercise.

For enterprises the dramatically increased pricing and complexity of App Engine’s new pricing model will become the cautionary tale to those pushing an enterprise to adopt a cloud offering until the pricing and pricing metrics are established.

Cloud leaders aim to control the entire technology stack
The second announcement, or lack thereof, that will affect cloud adoption is the news that “VMforce is dead”, to borrow words from Gartner analyst Yefim Natis.

A little over a year ago, and VMware made news by announcing a strategic alliance to let VMware and Spring developers build and deploy applications onto’s cloud platform.

Yefim broke the news about VMforce:

Yesterday at VMworld conference Tod Nielsen, a VMware executive leading its platform efforts, had announced that VMforce will not be delivered, CloudFoundry technology will not run in the data center and users of will be enabled to access in some unspecified way as a compensating feature. Today Byron Sebastian, platform executive, confirmed it. VMforce is dead.

Yefim repeats a long-standing Gartner maxim, “the only real partnerships are acquisitions”. went out an acquired Heroku to replace the VMware capabilities in VMforce.

Platform vendors, such as IBM, Microsoft, Oracle and SAP, control the entire technology stack underlying their platforms. As Yefim points out, this strategy will be replicated in the cloud arena. It’ll happen because cloud vendors, such as, are vying to join the ranks of platform vendors.

Enterprises and developers relying on cloud providers whose platforms are a collection of partnerships and strategic alliances are walking a slippery slope.

When these partnerships break down, developer and IT investments in applications that relied on these partnerships need to be migrated, rewritten or thrown away, resulting in wasted time, effort and money.

Enterprise developers and CIOs attending or paying attention to the news from VMworld or DreamForce 2011 have lots of exciting products and services to consider spending their time and money on. However, they’ll be weary of doing so without clear and long term reliable pricing and using platforms that a single vendor can deliver. This higher level of scrutiny is good for the cloud market, for clients and vendors.

I should state: “The postings on this site are my own and don’t necessarily represent IBM’s positions, strategies, or opinions.