SOA - Integration Industry Pulse

Beth Gold-Bernstein

Tracking SOA Transactions

user-pic
Vote 0 Votes

This week AmberPoint introduced its Business Transaction Safety Net™. This non-invasive solution monitors, records, provides search capabilities and can isolate root cause of a failure for end-to-end business transactions that cross applications and platforms. This release doesn't just track transactions - it allows real time visibility into the overall distributed application and can provide alerts or other actions in the case of failures.

Finding root cause of failures in distributed heterogeneous environments, where a transaction may involve multiple applications and platforms is a highly complex issue. While some tools enable you to pre-define criteria, Amberpoint now provides an automated way to search the log using technical or business criteria to discover the cause of the problem. The search capability enables operations to quickly figure out what went wrong. It supports policy based logging of all transaction history. You can define what to keep, what to encrypt, and even what to do with the data.

Tracking the cause of a failed transaction across distributed processes and platforms is a complex and time consuming process. Therefore, it's probably fairly easy to figure out an ROI for this solution. If you have the problem, and automated solution can save lots of time and money.

The key is having the right metrics to make that decision. That's also part of governance. While SOA governance is typically discussed in terms of design time and run time repositories, in a larger sense it's about clearly defining the roles, responsibilities, processes, policies, and metrics, and then monitoring and managing them, throughout the entire lifecycle, taking corrective action when needed. It requires aligning bonuses to overall success (something the financial industry especially needs to work on).

Hard times tend to make people and organizations marshal their resources more carefully, and spend on only what is necessary. When it comes to IT, how do you know what is absolutely necessary? What are the requirements, guidelines and metrics used? Who defines them? Who ensures they are followed? Once you get the roles, responsibilities, processes, policies and metrics right, there's a lot of great technology out there than can automate management for you. Success demands that you know how to use the technology effectively. How do you do all that in hard times? Pick your battles carefully. Only do what is necessary, but do ALL that is necessary. Effective governance is an absolute necessity.

1 Comment

Hi Beth,

Interesting, I agree that distributed failures can lead to trouble in diagnosing the cause.

Even more important than monitoring (IMHO) is managing, i.e., by preventing that distributed (partial) failures do not affect the validity of your business data. This is what my company does:

http://www.atomikos.com

Our products provide pro-active SOA transaction management. I think both approaches combined can be an even more powerful approach?

Cheers
Guy

Industry trends and vendor spotlights from Beth Gold-Bernstein.

Beth Gold-Bernstein

Beth Gold-Bernstein is a recognized expert in integration technologies and SOA with over 20 years experience View more

Subscribe

 Subscribe in a reader

Recently Commented On

Monthly Archives

Blogs

ADVERTISEMENT