Thursday, May 21, 2009

Is Enterprise ready for the Cloud?

This is one of the questions being bandied about in the Cloud world. And this is different than "is the Cloud ready for the enterprise".

I look to the System Integrator market as a guide. For example, Japan traditionally has been an environment where very little proprietary or custom application development is done by enterprises themselves, it is done by "SIers". The enterprises that have the core competencies to do custom development don't keep it in house, they in fact create an SI business. This is why you run into SIs named after large industrial manufacturers.

Increasingly this is occurring in the US and Western Europe; enterprises depend on SI's for implementation and even process and governance in some cases.

What does this mean? It means that whatever the internal IT budget is, many organizations have a greater appetite for the use of IT to enable their business than they are willing or able to put directly into the hands of their IT department. I take from this, that the limit is not available funds for the IT department, it is the IT department either cannot (or is perceived to be unable t0) effectively utilize the funds. The SI becomes one of the key levers business management and the CIO have to expand capacity.

Prior to cloud computing, the SI still, for the most part, had to come back to the IT department/data center operations "hat in hand" to get the implementation of the project deployed. Now the SI's will have multiple cloud relationships or run their own cloud. This creates another lever the enterprise. And, it grows the prominence and footprint of the SI in their business relationship with the enterprise.

It even begins to make them more of a channel for other vendors goods and services than they are today. I am thinking "go long" some of the key SIs in your portfolio as the market rebounds. A lot of them are sitting on their hands right now, actually investing less than my company CohesiveFT does, to meet customers needs in the cloud. But some of them look ready to move with real resources behind real business plans.

Wednesday, May 20, 2009

"What is a cloud" ruckus

Memo to everyone.
Dateline Chicago.

We are not going to end up with a uniform definition of "cloud".

We have been doing "open systems", "open standards" and "open source" for decades now - and we don't have a uniform definition of "open".

We have been "patching" software for decades and I can tell you there is no uniform definition around what a "patch" is, or what the behavior of "patching" actually entails.

We are developing a Rosetta stone of cloud-ish attributes. And I for one am ok with the Sun definition. I think Dave Douglas, Lew and company have done a good job of making it generally understandable without being overly proscriptive or prescriptive.

For my interests I have a subset of the Sun view.

Pat Cloud is:

Cloud User defines a workload
This is done manually or via automation. The workload is in the form of a "language module" (Java EAR, Java WAR, Python Egg, Ruby gem, Python-Django-AppEngine tar file).

Cloud User moves workload to cloud (either manually or via automation)

The workload uses resources
It dynamically accesses and consumes network, storage, processing, and possibly some value added services.

The Cloud has certain minimal attributes
It has an API.
It takes credit cards or PayPal.
It allows short time window usage (pay by the minute, hour, day, week?)

This works for me - and helps me get a lot of work done. Cheers.