Hopefully you saw the Intel and ODCA announcements today. The formation of an end-user organization to drive requirements into the industry through usage models is a long overdue development.

Also, I hope you’ve seen the Intel Cloud Builders reference architectures that we published on Wednesday October 27. It is an impressive set of reference architectures that go well beyond the IaaS usage model and into additional usage models including trusted pools and, policy based power management.

The development of these usage models into reference architectures really highlighted both the challenges as well as the dynamic nature of this segment of the industry.

There were a ton of decisions we made in the process of designing and building these reference architectures. Everything from the form of isolation for multi-tenancy, to the type of storage architecture to us, to use of trusted pools had to be specified and understood in the context of our overall objectives. Even though we had access to experts from across the industry, the process was definitely non-trivial. The challenge for an IT team is clear: gathering and learning from this huge variety is hard enough even when explained in the context of a usage model.

It really highlighted to me the value of using usage models as the starting point for these reference architectures. In that context, the ODCA is in a unique situation to challenge the industry to address their critical IT challenges. Having a set of sophisticated IT users express their challenges is a powerful statement.

In posts over the next few weeks, I will dig deeper into the reference architectures based on the set of initial usage models and starting with trusted pools.

Today, it seems like the majority of cloud computing discussion has focused on the data center. The end point, or client device, is often treated as more of an afterthought. My goal is to give you a different perspective on why, even for cloud based applications, the device you choose matters.

What does it mean to be Client Aware?

Simply put, "client aware" means that cloud based applications or services are able to recognize and take advantage of the capabilities of the end point device. Why does this matter? Organizations typically support broad range of applications. And many of these applications are becoming increasingly graphs and processor intensive. The growing need for security is also driving increased use of processor intensive encryption and decryption. One option is to rely solely on the cloud based data center to power the delivery of these applications. Another option is to optimize application delivery by taking advantage of the combined performance of the end point device and the cloud based to support these applications. By ensuring their cloud based applications are client aware, enterprises are able to take advantage of this more balanced approach.

The end point you choose matters

For cloud-based applications, it's important to look beyond availability metrics and consider the end-user experience. Most clouds are designed without any thought to the end point device. Yet, with larger screens, better graphics, stronger security and faster performance, some devices are clearly more capable than others. By taking advantage of the capabilities of the end point device, IT maintains a more flexible infrastructure with the ability to optimize application delivery and end user experience.

Making it happening

Intel is enabling this capability by delivering a series of web based APIs that enable applications to comprehend characteristics of the end point including network bandwidth, processor performance and available battery life.

For an overview of the topic, download the whitepaper on the benefits of a client-aware cloud.

To go deeper, download the APIs that enable client detection at http://software.intel.com/sites/whatif/webapis/.

To see how one application takes advantage a client aware infrastructure take a look at the video "Cloud Computing & The Web: Intel Web API for the Client Aware Cloud."

As you begin migrating applications to the cloud and want to make them client-aware, be sure to contact Intel with any questions. We'll be here to help.

Today is a big day for Intel, as we announced a major effort to enable next generation data centers for cloud computing. The topic of cloud computing is generating a ton of industry announcements, which is understandable given the once in a decade shift that cloud represents. Intel’s announcements are unique in the breadth of participation of the industry, as well as our support for a stronger voice of the customer. We announced our advisory role in the newly formed Open Data Center Alliance, a group of over 50 companies representing over $40B of IT spending who are committed to driving an interoperable, multi-vendor approach to the next generation data center. We also announced a dramatic expansion of our Cloud Builders program, with 20 different reference architectures from the leading providers of technology for cloud computing. All of this is in the context of our vision – a cloud that is federated, automated, and client-aware.

The announcement begs the question of what Intel is trying to achieve. After all, we are a key technology ingredient rather than a full provider of solutions, so addressing the challenges of the cloud is beyond our ability to impact alone. The Open Data Center Alliance is committed to multi-vendor solutions, including at the processor level, so our assistance to the group does not gain us a specific advantage. Let me shed some light on our motivation. First, we believe that cloud computing is still quite nascent. I enjoyed this posting, which highlights the risk of overstating the impact of cloud short term, while painting a good target. Our primary goal is to work with the industry to accelerate the cloud architecture transition.  Second, we are strong proponents of the model that has driven innovation and cost advantages in the standard high volume server market. This is a model where competition exists at every horizontal layer of the solution, and customers can select best of breed each generation. Here is a very thoughtful assessment of the vendor lock-in considerations in the shift to cloud. Our goal is to drive standards and implementations that enable differentiation, while limiting lock-in over several generations.

We don’t expect that our assistance to the Open Data Center Alliance, nor our collaboration with industry partners in the Cloud Builders program, will fundamentally change the end state of cloud computing. Rather, we’re focused on accelerating the outcome that would naturally evolve. A good analogy is the growth of WiFi. Back in 2002 the technology was expensive and the access to WiFi hotspots was quite rare. We were on a trajectory to see wireless adoption in notebooks occur at a glacial pace. Intel launched the Intel® Centrino® processor technology based on wireless networking standards, but augmented the product with a broad effort to enable WiFi hotspots and to improve the user experience linking to the hotspots. Net result, we saw wireless go from single digit percentage adoption to 90% penetration in three short years. Did we create WiFi? Of course not. Did we accelerate the benefits users wanted anyway by coalescing the industry of OEMs and service providers? You bet. We hope to have a similar effect on the data center.

The question of why we believe cloud computing is good for Intel is fodder for a future blog. A prominent analyst recently suggested that in just a few years the boom of the Intel server market would be a fading memory due in part to cloud computing. Throughout the history of the computing industry, we’ve had prognostications of an end to the innovation cycle. Meanwhile, we’ve seen the cost of computing drop, computing resources become ever more democratized, and new uses continually defy the imagination (or lack thereof) of the cynics. If both our products and our engagement with partners and customers drive an environment where computing is more capable and more accessible, our business results will take care of themselves.

I’d love to hear your thoughts – are you worried about interoperability, or more comfortable with a fully integrated solution?  Do you think cloud is all about reducing the footprint and investment required for IT, or are you excited by the new possibilities cloud computing creates?