We just published a new paper that explores the area of risk and cloud application optimization. Does it make sense to refactor that existing application? Should I make it run on the cloud or optimize it? What is cloud computing application utopia?

You can find it here…

https://www.sun.com/offers/details/cloud_refactoring.xml 

Table of Contents…

Benefits of cloud computing
Risks of cloud computing
Reducing risk
Definitions
Compatible with the cloud
Runs in the cloud
Optimized for the cloud
Patterns
Methodology
Process overview
Example: two-tier Web service
Initial assessment
Optimize through refactoring
Optimize further
Example: enterprise database cluster
Initial assessment
Optimize through refactoring further

See http://bit.ly/1MJoDW

I’ll be presenting at TSW 09 — Here’s the abstract…

DRAFT PRESO — here.
Carolan_TSW_09_Talk_04.pdf

Learn:

* An understanding of the impact of cloud-like infrastructures
* How we must change our support models
* How we design better products in a cloud-like world

Why You Need to Attend This Session:

In this economically challenging time, technology services providers must make decisions on how to best invest for the future while maintaining service levels today. Supporting services that are running on the cloud will be different than supporting those of the past. This session will help you balance this investment looking forward.

Presentation Abstract:

This session will discuss how dynamic infrastructure, cloud computing, and the increasing use of virtualization and automation impact support, monitoring, operations, and general serviceability of cloud-like infrastructure. As data centers continue to grow larger and become more complicated, what techniques can and should be used to see the important events and issues that impact service levels? Is that server going down “architecturally” significant or should it fail in place? How does this impact organizational issues and operations? What are some of the patterns being used to support large-scale dynamic systems and how can you implement them in your data center or support organization? How do we develop better products and services to help us see the “forest through the trees?”

So I’ve had the opportunity to think about what it means to provide “architectural” stickiness to customers recently. I’ll diverge a bit from my normal external view and look internally at how technology companies provide solutions to their customers — and how they help them “keep coming back” for more.

A quick plug, and some motivation to this article is my talk at the Technology Services World Fall 09 Conference in a couple weeks. I’m doing a talk on how the cloud has impacted services including service delivery.

So what do I mean by stickiness? Let me start with some examples — and networking technology companies know this well — being in the traffic path day in and day out for your customer’s customers most important transactions mean that customers know who their critical technology providers are, and the network is usually one of the most important elements.

I’m calling this architectural because of this reason. If you simply play in one tier of the environment (e.g. let’s say DB) then you’ll only have an opportunity to talk to your customers when they need to modify or add something around or to that tier. Sure, it could be “business critical” but often new, innovative ideas might not be so critical when they start out.

Other examples of architectural stickiness…

Product implementations that span data center tiers — e.g. networking
(discussed above) architectures and products — again in the data path
– security products as they may inspect application characteristics and flows, auth
– virtualization technologies — especially ones that exploit the operational lifecycle

Management products — which is an expanding space

— active deployment and management products such as provisioning tools, that add value, consistency, integrity, etc.
—-I’m not sure that I would include a patch manager here?
- there’s an element here thats active — taking standards and translating them into reality — like what N1 Service Provisioning System (N1 SPS) and Opsware do. They sit at the forefront of programmatically effecting deployments. They are hard to remove once engrained.

other standards – e.g. Architectural standards and designs
— these too can be sticky as much as they are referenced and used.

The elements above all have active vs passive elements. Active data path, authorization, deployment.

How do you help your customers come back to you?

Technorati Tags:
cloud computing, data center architecture, networking, automation

See my other blog postings at http://blogs.sun.com/jasoncatsun

Just a bit about me:

I’m an engineer and architect at Sun Microsystems, helping Sun customers understand, design, and implement technologies around data center architecture, automation, cloud computing, and the integration of all of this as next-generation networked systems. I take much of my influence from not the trade press but research and cross-sections of science and technologies including biologically inspired computing, ecology, large-scale networked systems used by the military, and other areas.

I very much believe in Mark Weiser’s vision of ubiquitous computing. The 80/90’s had the rise of the PC — 1990/2000s had the rise of the internet — now we have the rise of both, together, as systems designed with networking and processing together beyond the desktop to everything we do and experience.

My daughter did not need to take a class to understand how to use the iphone — its a long ways forward from the CPM/DOS/Unix days of the past 30 years. What’s next?