Richard Sykes3

According to Wikipedia, a polemic is a contentious argument that is intended to establish the truth of a specific belief and the falsity of the contrary belief. I am sufficiently concerned about certain current aspects of the business development of the cloud to polemicise on the matter.

I wrote last month about Cloud Expo Europe 2013 whose marketing message was 'Unlock the Cloud'. I noted that the vast majority of the exhibitors at the Expo were not in software services and apps businesses, but in infrastructure, platform and related businesses. The Expo was really about the'tin': we are still at the early construction stage of the cloud and we need to remember that.

So to my key issue No. 1. At the core of the cloud is the 'tin': data centres and server farms which are tightly network-integrated and designed to operate through exploitation of virtualisation to deliver a high degree of automation, very high asset efficiency, reliability measured in the multiple 9s, and the flexibility to turn an ocean liner on a sixpence in microseconds. I am talking about the likes of the Amazon Elastic Cloud and Google Cloud, both of which set today's competitive benchmarks. I am talking about operational design and practice that draws on the accumulation of years of intense learning driven by operating in highly competitive consumer markets.

This is why the private cloud will not prove a sustainable reality, let alone a sensible option. The design and sheer operational experience integral to the Amazon and Google Clouds gives them major competitive leeway - even a major bank or a massive volume retailer will have difficulty reinventing what Amazon and Google now offer routinely to the market. In the words of Jim Darragh of Abiquo: "The operation of these private clouds are at the side of the business, not at its heart". Such in-house operations will not cut the mustard.

So I refer you to my earlier Weetabix/Wolseley business analysis. At the operational core of the cloud the winners will be in the Weetabix business: highly standardised services manufactured in highly automated factories operated by management whose sole focus is in manufacturing and distributing services. Their intellectual challenge lies in manfacuring excellence. It is all they talk about, day and night. This is commodity business made very profitable.

If your intended IaaS/PaaS supplier does not fit this description stop your procurement there! This will rule out half those vendors claiming to be cloud IaaS/PaaS suppliers, and rightly so. This is where HP got its acquisition of EDS expensively wrong. It failed to use the opportunity to create a head-on, standardised and secure competitor to the Amazon Cloud in the enterprise space.

On to my key issue No. 2. A company I know well is developing an IT operational model that requires an integration/middleware capability. As they are migrating to the cloud they sought to source this capability as a cloud service. Market research identified four major players all of whom, according to the marketing messages on their web sites, offered this capability as a cloud service. Workshops were held between the company and each of the four to define the scope of the integration/middleware services to be procured. A procurement tender was then released, but all four vendors declined to bid - their clearly marketed cloud services proved, on closer examination, to be vapourware.

I hear versions of this story quite regularly. The marketing imperative of the day is the cloud, so the message to the board is: "We must be seen to be in the cloud". Capabilities from the old world are rapidly re-dressed to appear as though they are of the new.

But of course they are not. Vendors should only market cloud services that are operationally proven. If they are at proof-of-concept stage they should say so publicly and seek clients willing to invest in several years of co-development.

Last month I applauded Dell's move to take itself off the public markets so as to create the space to re-engineer its fundamental business models. Other vendors need to follow suite. They have to make the key decision - is their future as a Weetabix or a Wolselely? That decision made, they need to quickly put in place the very major restructuring that will undoubtedly be required.

What they cannot do is market new-generation cloud services that on closer examination turn out to be vapourware. That, surely, is the road to perdition!