Dealing with Performance Challenges of IT systems for ‘Load Burst’ Scenarios

Of late, while trying to adapt to digital wallets for bill payments and online shopping, I came across promotional offers across newspapers, media and TV channels for one of the popular digital wallets in India. Thinking to get myself benefitted from those promotions, tried to open up their Mobile App and found it unresponsive with multiple error messages – it could not even properly display the balance in my wallet. I switched to their website from my laptop and came across ‘Service Unavailable’ error. In the interest of time, I proceeded with my shopping and bill payment needs with other wallet providers. What might have gone wrong right after the promotional offers, I wondered!!! I checked their App and website intermittently throughout the day however I could see the App and website restored to normalcy only the next day. Presumed that a sudden surge of end-users to their servers due to Promotional Offers might have caused the unresponsiveness and unprecedented downtime – which I am sure might have led to loss of business as well.

Similar downtimes reported by one of the major UK retailer during 2016 Black Friday – the online shoppers reported that the website was ‘down’ once the Black Friday deals were live on midnight. End users reported sluggish behavior in Social Media as well in order to get quicker attention and resolution from the retailer.

With this context, I would like to highlight on ‘Load Bursts’, its impact on IT systems’ performance & scalability, key challenges, various aspects that should be considered to effectively deal with so that organizations do not suffer from any business loss due to system’s performance & scalability issues caused by sudden load surges or load bursts.

Load Burst can be defined as “a sudden increase in user or transaction activity on IT systems infrastructure (such as webservers, application servers, database servers, middleware servers to name a few) that typically lasts for a short span of time”. It will be typically in the order of millions (100K, 200K users or transactions), which can seriously impact businesses if not addressed proactively. Though the surge can be for a very short duration, the impact it causes on application can last for several hours. Load Burst scenario typically occurs during holiday season, new product launches, marketing campaigns / promotional events, application roll-outs to new regions, application consolidation due to mergers & acquisitions, to name a few.

With digitization picking up fast in every walk of our life, mobile and online has become the preferred channel of interaction for every end-user for daily needs such as buying groceries to investing in stock markets. Also, with businesses expanding globally to cater to users across geographies, the end-user base and traffic to their IT systems is increasing multi-fold which is one of the key factors to cause Load Burst situations. Hence, it is quite important that websites, mobile sites, mobile ‘Apps’, Products/Services, Point of Sale systems, back office systems should be highly responsive, 24*7 available with zero downtime so that they can handle Load Surges effectively and provide a better user experience, which is a key for the sustenance of any business in today’s competitive world. So, how organizations can be ready to handle such scenarios to avoid business impact and reputation loss?

The answer is as simple as ‘being proactive to assess & validate IT systems’ performance for burst loads’ and identify ‘weak-links’ in application stack so that necessary application tuning along with any additional infrastructure provisioning can be planned effectively. Setting aside the technicalities required to address the challenges of Load Bursts, I would like to clearly highlight that a mindset change amongst the Business Owners and IT Managers is the key to focus on application’s non-functional requirements in parallel to the feature-rich functional requirements. Most of the times, performance is an after-thought compared to developing features which typically causes lesser focus for application performance & scalability for anticipated production workloads. A proactive focus & tracking of non-functional requirements w.r.t end-users, transaction volumes, anticipated performance SLAs and infrastructure capacities right from requirement gathering ensures.

With this, I would like to leave a thought to ponder if addressing performance challenges for load burst scenarios for IT systems is any different compared to traditional performance testing – I will be back with the details in the next blog.

About the Author

Madhu Tanikella is a Senior Technology Architect at Infosys, Hyderabad.  He has 15+ years of experience in providing delivery and consulting services in Performance Engineering & Performance Tuning for enterprise software systems that fall under client-server, multi-tier (Java/J2EE), SOA, Messaging Middleware architectures, Virtualization and Cloud models and custom-built applications with multiple COTS products. He has specialized expertise in the areas of NFR & SLA Validation, Workload Modeling, Performance Modeling, Bottleneck Identification and Performance Tuning (of Java/J2EE, Oracle DB, TIBCO Business Works and IBM MQ), Capacity Planning, Hardware Assessment and Sizing for distributed software systems. He also has expertise in assessing performance practices in an organization and setting up Performance Engineering CoEs for various clients in different Domains such as Financial Services, Manufacturing, Insurance, Healthcare, Retail etc. Madhu Tanikella is certified in IBM Cloud Computing Architecture and IBM SOA Solution Designer. He can be reached at [email protected]