3 ways to achieve and ensure application performance following a rollout

Rolling out a new critical application such as an ERP system is a vital objective for a CIO! There is a lot at stake from both a financial investment and a visibility standpoint. Any malfunction during the application rollout phase may have an impact on user adoption as well as slow down the business benefits of the new application. Getting things right from Day 1 is therefore critical to the IT team’s success. So, how can the infrastructure team contribute to that success?

Application rollouts — what’s the worst that can happen

Many things can go wrong when deploying an application. The level of visibility you have during the preparation and the rollout phase will indicate how quickly you will be able to identify an issue and fix it. It will then drive your ability to deliver the application on time and with success.

What are some things that you want to avoid?

  • Application errors
  • Slow application performance, including abnormally high response times
  • Poor network delivery

And above all, you want to avoid a situation in which end-users report this—and you cannot quickly identify the root cause of the problem and repair it within a few hours.

How can you be prepared?

Easy. The three Ps:

  • Plan
  • Prepare
  • Perform

Well, the very first things you probably want to do as an infrastructure / IT operations team are:

  • Be aware of the application rollout well in advance,
  • Be involved in the testing and pre-production phases,
  • Get enough visibility to prepare for an adequate application delivery architecture,
  • Have enough visibility to quickly determine what goes wrong (in case something happens) during application deployment.

Let’s review the key criteria / data you need to go through for that preparation.

1. Understand the profile and drivers of application performance

the evolution of application performance over time

Even before your application starts to be used, there are some pieces of information that you will want to collect to ensure that you will be able to deliver it to your end-users:

  • Network capacity required
    • For each end-user, what is the bandwidth required to access the application?
    • Does the bandwidth vary depending on the end-user’s profile?
    • Do we have enough bandwidth in place to ensure sufficient application performance?
  • Application performance drivers
    • What drives the end-user experience?
    • What portion of the transaction is related to server processing vs. data transfers?
    • By what factor can network performance improve/influence application delivery?

These application performance drivers can help you to:

  • Size your network appropriately
  • Identify the right application delivery optimization solutions (and avoid wasting time and money on those which will not be helpful)

Depending on the data you gather, you may then want to investigate different options such as:

  • Do “nothing” (if nothing is required and it saves your IT teams some money)
  • Invest in application optimization (optimizing the server processing time and response size)
  • Invest in network optimization (compression, prioritization)
  • Find alternative ways to access the application (e.g. thin client)

2. Identify and fix defects and bottlenecks in pre-production

In pre-production, you need to monitor for the following phenomenons:

  • Slow front end response
    • When is the front server responding slower?
    • For which transactions and queries do you get slow responses?
  • Application errors
    • For which transactions do you get application errors?
  • Backend bottlenecks and slow processing
    • You will probably want to also monitor the back end transactions (database, file servers, etc.).

By having visibility on all this, you will be able to share application performance issues, build a fix, and validate the proper execution of the application before it becomes available to all end-users.

3. Monitor the end-user experience in the ramp-up phase

The behavior of the application may vary depending upon:

  • The load
  • The location of end-users
  • The connectivity between end-users and data centers.

It is therefore crucial to keep the application under control and monitor the end-user experience during the ramp-up phase in order to:

  • Identify sizing issues appearing at a given load
  • Identify issues which are specific to a given location

Thanks to this continuous monitoring, you will be able to keep the resolution times short and sustain the trust the end-users set in the application. As well as the trust that your team stores in its collaboration with the applications team.

In fact, wire data can help you roll out your applications successfully.