Darren Williams

Prescriptions in IT: Treat the cause, not the symptom

July 02, 2013 6:00 AM EDT

Maybe a server is temporarily down, or perhaps an end user is unable to log onto a device due to password failure. But what if that server is an integral part of a manufacturing workflow? Or what if that blocked login meant that an entire team of employees was unable to obtain their work orders for the day? 

It can be the most seemingly innocuous of circumstances that can lead to catastrophic results for a business. Maybe the first sign is an end user filing an incident because their email isn’t working properly. On the face of it, this is a fairly simple request. The IT technician can reference and provide standard fixes that should solve the immediately apparent problem.

But what if the IT technician is looking at a symptom rather than the cause? What if the email account is linked to a server that also supports the order management part of the business? And what if the server has crashed? What if, for every minute that the system is down, orders are not being processed, customers are becoming frustrated, and, ultimately, business is being lost because a crucial part of the infrastructure has collapsed. 

Data driven decisions

As with most things in technology, it really boils down to the data.  In the above email scenario, the next step will depend upon the type of insight available to the technician, and this will rely 100 percent on the data and technology that is available to him or her at the time the service incident is filed.

If all that technician sees is the first layer of data – a ticket that’s been filed by the end user – then this will be logged as an email problem and that is what they’ll fix. The technician will reference the email trouble-shooting FAQ, provide the necessary information to the user, and move on to the next incident.

But, as many technicians can attest, an epic failure can often masquerade initially as a typical “day in the life of IT” scenario. It’s not until that IT person is well down the path of remediation that the full impact of the problem on the business is uncovered. Hindsight, as we always say, is 20/20.

Following data's path

In order to make a comprehensive diagnosis, this is exactly what the technician needs to do. Follow the paths their data provides to compile a complete picture of the complex layers that comprise their organization’s infrastructure. This will allow them to trace an issue down to the root cause, so they can assess (and ultimately avoid) other related impacts to the business.

It’s like a patient visiting five times in a row with a persistent cough and minor fever; the doctor can assume it’s just a cold and prescribe fluids and rest and send them on their way. A good doctor, however, will take a closer examination, ask the right questions and order tests, digging into the matter to ensure that the symptoms map to the appropriate cause. One of my favorite medical sayings is, “when you hear hoof beats, think zebras,” and sometimes this is what IT needs to do.

However, if the infrastructure available to them is only skin-deep, providing IT managers with a basic layer of information – then break out the cough syrup and hope the best for your business.

Tying it together with technology

Successful outcomes in IT happen when the IT organization is telepathically quick and effective at restoring services in the event of failures and outages. The key to their success though is access to a complete picture of the infrastructure they support, recorded in terms relevant to the organization's requirements.

The necessary data to build this picture already exists throughout the organization, but in order for it to deliver any value, it needs to be collected and organized so that IT can access a data-driven view of the entire business.

Typical of most organizations, hardware and software information is stored in multiple repositories with distributed configuration item (CI) information residing on external management tools. These management tools span multiple corporate domains, including asset management and event management applications and, thus, capture different information for the same CI.

In order to approach each service incident deeply and proactively to get to the root of things, IT needs a way to look at this disparate CI information all at once.  This is where IT Service Management technology can really make a difference in a business, providing organizations with the ability to automatically consolidate CI information from a variety of internal tools within a Configuration Management Database or CMDB.

The CMDB allows service organizations to view information from independent sources for a more holistic view of the IT infrastructure and their relationships.

Preemptive insight

With such deep insights into how the business is connected, IT managers can even start diagnosing problems before they begin. Effective ITSM solutions will include an integration engine, ideally one that uses XML descriptor files embedded within the application to connect to foreign hosts. Based on the details entered, asset management data can be transferred to a staging area where it is assessed against category maps and merged in a CI within the CMDB. 

This means that any future syncs of the integration engine will update the CIs directly, resulting in less work for IT, or generate a change request to alert IT that a potential issue need be addressed before deployment. This alerting process can occur hours, even days, before the first service request is filed by an end user who encounters a symptom (such as email failure) of a more systemic cause, allowing IT to predict and avoid significant impact to the business.

Talk to your IT service organization and determine the depth of view they possess.  If necessary, invest in technology that will allow them to harness the data that already resides within the business to truly diagnose and preempt problems. This will allow them to convert potentially epic failures into timely but standardized IT responses, focused on the cause versus the symptom.