Intel badges the cloud

There was a time when Intel’s marketing tactic of badging PCs and the like with the famous `Intel Inside’ tag made both good marketing and commercial sense. One can see, therefore, that attempting to do it again, now that the cloud is becoming the dominant technological underpinning to the exploitation of IT resources by businesses around the world, would seem to make good sense again.

But does it? Indeed, maybe the question really is: `why on earth should it make good sense now that the cloud has shifted the emphasis, from the users’ point of view, away from technology per se and on to delivery of services and just what those services can achieve for those users?’.

First of all, here are the basics of what Intel is announcing. The company has unveiled the `Powered by Intel Cloud Technology’ badge. It has also teamed up with 16 leading Cloud Service Providers (CSPs), to exploit the badge to provide cloud users with a clear view into the technology powering a CSP’s infrastructure before they purchase any services or instances.

The program benefits CSPs by communicating differentiation in service offerings and informs customers how infrastructure technology can improve service performance, reliability and security and help them maximise Return on Investment.

It also provides the Intel Cloud Finder online search engine, which is designed to help users find CSPs that provide cloud services built on Intel technologies that match required criteria, with the ability to try services for free before purchase.

In Intel’s view, as IaaS revenues continue to grow by an estimated 41 percent annually by 2016, and more companies consider outsourcing their IT services, the technology that powers cloud-based services and applications will matter more than ever. CSPs participating in the new Intel program will be recognised by using the `Powered by Intel Cloud Technology’ badge to distinguish their Intel-based instances where the performance and security capabilities of the underlying hardware become transparent to the end user.   

The 16 CSPs so far recruited are Canopy (U.K.); Virtustream (U.S.);  Cloud4com (Czech Republic); CloudWatt (France), Expedient (U.S.); KIO Networks (Mexico); KT (Korea); LocaWeb (Brazil); NxtGen (India); Online.net (France); OVH (France); Rackspace (U.S.); Savvis (US); Selectel (Russia); and UOLhost (Brazil). These follow on from the original collaboration between Intel and Amazon Web Services last September.

Some of Intel’s technical argument behind the badging move is a research paper, `Exploiting Hardware Heterogeneity within the Same Instance Type of Amazon EC2’, written by Zhonghong Ou, Hao Zhuang, Jukka K. Nurminen, and Antti Yl¨a-J¨a¨aski, of Aalto University, Finland; and Pan Hui Deutsch Telekom Laboratories, Germany. This suggests that a heterogeneous cloud infrastructure environment may result in 40 to 60 percent performance variation.

Therefore, end users are increasingly looking for more insight into the performance, capabilities and cost trade-offs of the many instances that CSPs offer so they can get the right size and type of performance matched to their specific workloads.

The company suggests even a slight change in cloud instance performance can impact sales on e-commerce sites. It quotes one retailer, Shopzilla, increasing its revenue by up to 12 percent by accelerating its page load times from seven to two seconds. For enterprise and other IT buyers evaluating hybrid or public cloud services, CSPs participating in the program will provide detailed information about available CPUs, hardware acceleration features, storage, software and network capabilities – all of which can have a have significant impact on applications, costs and experiences for end users.

It is possible to suggest, however, that hardware performance issues are only part of the overall performance position for a CSP, and arguably a relatively small part when it comes to the end results of cloud service delivery. The design of applications software, service management tools, service orchestration and the rest – and the developer’s and orchestrator’s understanding of the interaction and collaboration between those elements at an API level are far more important.

For example, with the retail issue cited, page loading speed can certainly be a contributor – though much more of that is geared to page design issues, where retailers simply load content into webpages from product suppliers’ own websites from around the world – and the page won’t load until the last contributed element of contributed content is in place. Its arrival, however, is subject to many more variables than simply processor/server performance.

Also, more revenue could be gained by better management of such websites. For example, the use of new technologies such as the vAC Virtual Access Control system from UK start up, NetPrecept, can help retailers identify those website visitors who are likely to move to making purchases, rather than just `tyre-kicking’. This can be vital when website capacity limits are reached, for it provides tools to help manage the latter group, for example by offering them preferential service vouchers to come back later.

These are software tools developments and techniques which are far more effective than just using a type of processor or server. And then again, those servers are by and large commodity items, and it is highly likely that most CSPs would not necessarily know whether an individual server used is an Intel processor, or a functionally equivalent AMD device.

Lastly, and perhaps most important, the cloud provides the opportunity for users and CSPs to offer service diversity through system diversity – and even have those diverse systems collaborate at an API level. There is no reason why a service provider should not offer a specialist hardware capability such a powerful Oracle Exabyte machine that can be loosely coupled with commodity servers to provide the rich service level that a specific customer might require.

Even more important here is the fact that there is now a growing number of server vendors starting to offer systems that use processors based on ARM technologies. Given that ARM-based processor designs are endemic in the smartphone and tablet world, where Intel plays the part of also-ran, and that the world of smartphones, tablets and the cloud are becoming ever-more closely linked and entwined, having ARM-based servers is becoming as strong an argument as having an x86-family server, let alone an Intel-specific one.

So from Intel’s perspective yes, the new badge is a good option, if only in terms of perception marketing. For the business user of cloud services, however, there has to be some doubt as to its practical relevance this time around.