Design and implementation of systems in accordance with client SLA’s, and why you should care.
In our Application Performance Management (APM) whitepaper, we mentioned that the key performance metric of any application is the end user experience.
Application performance can be measured in a number of ways. However, it is the impact on end users that has most relevance to businesses.
Services consumed by end users are the culmination of many IT components, supported by numerous IT groups within a business.
The “application” is delivered by networks, mainframes, storage servers, application servers, web servers and multiple endpoints and interfaces such as browsers, ATMs, kiosks, laptops, netbooks, tablets and smartphones. Therefore, if a service slows down, or stops working altogether, there are a whole host of potential causes and IT owner/ resolver areas to be explored. Good APM enables end to end management and visibility of a customer-facing service: from datacentre to end user.
Most CEOs are not conversant with processor cycles and application server technology. However, they will be well aware of any impact on customers or profitability if a service goes down.
Common examples of performance issues include: a deluge of calls to the service desk from colleagues frustrated by increased desktop login times; slower response times on CRM systems impacting the productivity of call centre staff; or the more recent example of the Amadeus online flight booking application, which suffered a
3 hour outage. This impacted travellers trying to access the booking page on the British Airways website.
The Amadeus outage was suspected to have been caused by a “network component failure” and it appears to have been rectified relatively quickly.
However, in
September 2010, a 21 hour outage suffered by IT supplier, Navitaire, left thousands of Australian airline passengers stranded. The resulting disruption led to Virgin Blue providing food and overnight accommodations for affected passengers. That’s a tangible application performance impact that any CEO would take notice of.
Although the fault was isolated to a solid state disk server in its datacentre, used to host Virgin Blue’s travel booking service, Navitaire delayed the failover process and opted instead to try to repair the device. As a result, Navitaire was found to have breached its Service Level Agreements (SLAs) with Virgin Blue.
When designing a customer-facing service, it is vital to specify the SLAs associated with the components of application performance. For example, the amount of downtime acceptable to the customer base without impacting business. This should influence the system design at all layers, from the network up to the presentation of the service on end users’ screens.
The SLA should encompass the testing approach and system up time assurance, fault tolerance and failover mechanisms, along with methods of accessing the service and upgrade processes.
The important point to remember is that end user experience is the key metric by which an application’s performance must be gauged. Your datacentre, device and network monitoring may be showing normal levels of service and throughput. However, if booking agents are unable to process passengers’ flight details, then this is the real indicator that your application performance has degraded to unacceptable levels.
By implementing APM an organisation gains an integrated view of the application delivery. This allows them to be alerted to performance issues, before they affect the end user experience. In the event of end users experiencing degradation of service, this overall visibility of the application delivery system will enable organisations to quickly isolate the IT domains that are responsible.
In addition to protecting the company’s reputation with customers and maintaining brand loyalty, the ability to quickly identify faults and reduce mean time to recovery (MTTR) also helps service providers to avoid SLA penalties.
C&C provides diagnostic services using end to end monitoring that set up and capture metrics from the different application and infrastructure components. This process provides visibility into the performance of organisations’ applications and services from the data centre to the end user and provides fast results to resolve performance issues.
Before an organisation rolls out a new service, C&C will work with IT staff to profile applications and transactions and model performance against varied network conditions, peak demand periods and loadings. IT service managers can then set realistic service performance expectations and optimise performance prior to deployment. Once the performance baselines have been set for each stage of the service delivery and the network footprint is mapped out, C&C helps organisations to predict performance against a series of “what if” scenarios. This stress test modelling helps to ensure that there is sufficient capacity and redundancy built into the network to ensure seamless application delivery, while also allowing for future growth in the service.
APM allows organisations to step away from the break/fix mentality and to move to an ongoing application lifecycle management process that pre-empts and minimises any impact on end user experience.
To download the C&C Application Performance Management whitepaper, click
here
