Internet-Scale BSM

| 2 Comments | No TrackBacks

Colleague and friend Annie Shum shared with me fascinating data from her research on Cloud Computing. According to Annie, the economics of mega datacenters are compelling:

The study concludes that hosted services by Cloud providers with super large datacenters (at least tens of thousands of servers) can achieve enormous economies of scale of five to seven times over smaller scale (thousands of servers) medium deployments. The significant cost savings is driven primarily by scale.

In the context of BSM Review, the obvious question this study poses is the tweaking of Business Service Management to respond to and cope with operational and business challenges on such a scale. For example, at smaller scale configuration drift might be laboriously manageable through traditional techniques. For super large datacenters, however, it is a compound problem:

  • Exception handling is prohibitively expensive at large scale.
  • Scale economics are likely to diminish (due to configuration drift problems). 
  • The associated risk could be lethal. Large scale configuration drift might go beyond loss of data in an IT department - the datacenter operator might lose customer’s data.

Knowing Annie, I have no doubt she will elaborate at length and depth in this blog on various Cloud Computing aspects of Business Service Management such as Virtualization Sprawl. (See her recent article A measured Approach To Cloud Computing: Capacity Planning and Performance Assurance for the first “installment” on this important topic).   I will do the same with respect to Agile Business Service Management at grand scale. For example, an intriguing question is the setting, modus, operation and governance of the Application Support Team in this kind of environment. One can actually view it as a Venn diagram:

  • Cloud Operations on one ‘circle’ 
  • Customer Application Development on another
  • Application Support in the intersection 

Stay tuned!

No TrackBacks

TrackBack URL:


I have been dealing with "scale out BSM" extensively over the past few months within extremely large enterprises. What I am seeing is the need for broad based scale out of what I call "Generic BSM" or "BSM Lite" that can provide the solid foundation needed for operations and support teams to manage their environments which are not optimized as the "Internet 10" companies are (one support/admin for 10K+++ systems).

I've talked about this at length on my blog previously, but at scale out, they need "just enough" of the basics - application discovery/dependency tracking, proactive change drift notification, core fundamental instrumentation, monitoring and eventing and simple presentation layers that provide some basic context to move the ball forward to quicker triage, isolation and resolution.

Most extremely large enterprises I've seen struggle to keep up. The "gee whiz" features of software solutions aren't being looked at as near term needs - just give me something to solve my immediate pain points, to catch up, to stay ahead of the avalanche of changes. Once that happens, and some automation, simplification and standardization has been introduced then and only then can these companies move into those sexier things like fancy dashboards and what not.

The opportunities for scale out BSM are in the areas of what I call Hybrid BSM. Where simple, consolidated monitoring capabilities are combined with the tenants of BSM in a common, highly scalable solution that places little to no administrative burden on the teams who are already overworked.


Leave a comment


Type the characters you see in the picture above.

About this Entry

This page contains a single entry by Israel Gat published on November 10, 2009 5:47 AM.

Bad IT = Bad CEO? was the previous entry in this blog.

Peter J. McGarahan on Business Service Leadership [Part 1] is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.