CA Community






This Blog

Can SLM Be Substituted by Other ITIL Processes?

Published: March 11 2010, 10:15 AM
by Michael King

Sitting in LAX Airport about a week ago, waiting to see if the snow storm that was hammering the Northeast would cut me and other travelers a break and let us fly into the region, gave me lots of time to think about a customer I had just seen.  As I sat there, I reviewed and replayed the events in my mind and decided to write a long delayed blog entry about some of the revelations that came from these events.

While I was in California, I had the opportunity to speak to this customer about Service Management.  The discussion surrounded their implementation of ITIL, specifically Service Management.  During the discussion, the question of reporting SLA results came up.  The customer wanted to display SLA results and service statuses in a dashboard for various group leads so they would have a real-time view of services.  As we drilled down deeper into the requirements, we identified the desire for real-time results to be displayed.  Once I heard this, I realized that this need could be filled from two different ways: The SLA status and results could be managed using a SLM tool (SLM process) and the real-time QoS status using Network/Application Management tools (Availability and Capacity Management processes). 

Here is my reasoning: SLA Management tools are restricted in the real-time processing department.  They are dependent upon the frequency of data being delivered to them.  When measuring/monitoring SLAs, a balance needs to be struck so that the frequency of data collection does not negatively impact the service that is being monitored; simply put, collecting data so frequently can stress the service being measured and negatively impact the SLA.  If the goal of the real-time QoS monitoring is purely for Operations purposes, the IT organization can leverage its network systems management tool to manage this.  Some systems management tools afford the ability to model a service similar to that modeled within the SLM tool.  This feature allows IT organizations the ability to monitor QoS from a proactive and reactive perspective.

Another thing that I gained from this conversation was that there is sometimes a misconception that Availability Management can act in place of Service Level Management.  In some cases, Availability Management can be a close substitute for SLM; not always a replacement.  If the tool used for Availability Management gives the IT organization the ability to:

  • Model a service based upon the relationship of its constituent components
  • Apply business or service hours
  • Generate reports on a scheduled basis
  • Associate the report results to business and financial objectives

...then this tool should be able to replace an SLM tool and satisfy the reporting component of the SLM process.

As I look back on this, I am reminded that when creating and implementing a SLM process, remember that the other ITIL processes - Availability Management, Capacity Management, and Incident and Problem Management - all feed the SLM process with data points and can act as close substitutes to SLM but cannot replace SLM.  The SLM process is made solid when all of these ITIL processes are integrated and working together.

Share this post:  EmailEmail

 

By: Michael King
Michael King is a Senior Engineering Services Architect in CA’s Service Management group. Michael has over 19 years of experience in IT that includes software engineering, operations management, systems integrations, and process reengineering. Currently, Michael concentrates on Service Level Management...
Read More..

1 person has left a comment:

It seems this customer says "SLM" but is really focusing more narrowly on pieces of it like monitoring Service Level Objective (SLO) and Operational Level Agreement (OLA).

Another view (ITIL's) sees SLM as part of service design process for defining the service and its quality requirements from a customer perspective before it is put into operation (this should guide how the service is built, and avoid under-delivering and over-delivering).  

The word "agreement" in SLA is key: is there a documented agreement between the provider and customer?  If not, is it really an SLA?  If so, then the operational instrumentation and thresholds should be guided by the SLOs and OLAs modeled in the SLA.  The SLA monitoring/reporting doesn't need to be real-time, but if the metrics and thresholds managed by real-time QoS monitoring tools were set in accordance with the SLA, then the SLM process is doing a big part of its job.  

As for the idea that monitoring itself can slow the service down, this too may be a good conversation to have with the customer before the service is deployed.  How important is QoS to this customer?  Are they willing to pay an extra cost to build the service in a way that it can be monitored according to their quality requirements?  

Posted by: David Wilt | April 1, 2010 12:18 PM

Leave a Comment

* An asterisk indicates a required field

* :  

:

* :  

 Submit