Designing Data Centre automation solutions

Today was planned as a look at Sun’s system management solutions and the day was started by Sohrab Modi, the V/P for the group. We really have a problem with our branding, but we’re sticking with N1, managing maNy systems as 1. Sohrab’s presentation called “Simplify, Integrate, Automate”, hit two key points for me. The first is that its a solution, the second that there are two dimensions to virtualisation.

First, N1 is a solutions sale and the customer will define the scope and boundaries. Sun’s field need to capture these requirements and map both Sun & partner products to the requirements and take integration responsibility, at the least the solution design needs to be a collaborative process between the customer and vendor. Our proposition needs to be we meet your need and solve your problems. Its not good enough to try and win feature/benefit beauty parades. Sohrab showed how in the next 12 months Sun (his people) will be bringing virtualisation and management products to market to allow us to more address more and more comprehensive requirements. He demo’d some of the more advanced features of our logical domaining and provisioning technology which gives me confidence that his team are finally getting their delivery train in order, its been a long hard struggle for them. (One of his staff produced one of the analyst slides showing Sun in a strong position today, at least we’re gaining mind share.)

The second piece of his talk that I liked is his two dimensions of consolidation. The virtualisation technologies mean that the data centre might have more operating system images, which without automation is more difficult and costly to manage, therefore we have to aggregate systems into run time resources and management objects.


This something I have been calling pools, but I suspect that Sun and the industry will call them something else. The demand for multi-node pools by applications is a crucial part of the capability supply and demand nexus and the economics and architectural constraints that aggregation enables and and creates make solutions design and integration more important not less. It is unlikely that one technology can answer the management consolidation and virtualisation problems.

As said N1 is a solutions sell, and customers will define the boundaries of scope, and we need to meet all of the customer problem. Sohrab gets this and is committed to a full technology offering that meets customer needs. We’re not looking at point product releases, we’re selling/building a platform and field people i.e. me need to learn how to ensure that customers define problems, sun offer’s technology and the sales process collaborates on solutions design. As I said we don’t want to sell each of our N1 products on a feature benefit basis, and customers shouldn’t want to buy that way.

Presentations later in the day emphasised Sun’s co-opetive approach, building from the ground up to enable integrators to add value. These integrators are as likely to be customer internal staff as they are to be consultants. This means that a central piece of our platform is going to be the orchestration functionality so the “N1 backplane” will be able to orchestrate micro transactions, which themselves can be undertaken by either Sun product or partner/competitor product. It all enables solutions design.

The full presentation is still available here….


Originally posted on my sun/oracle blog, republished here in March 2016


  1. This initiative failed. Sun’s reticence to develop a real VM hypervisor, its legacy of big iron sales and its sectarianism around Solaris were all problems that the company never overcame.