digital abstract

Concurrency Is Hard Currency In Business Apps

digital abstractBuild and they shall run – non-stop

So how often does your company update its business apps? Software update intervals are clearly getting shorter and shorter as our online, and especially mobile, activity levels rise. This is happening concurrently on several fronts: the inter-personal communications; the e-commerce activities; and the exploding field of IoT (Internet of Things) communication. This speed-up reflects four major trends: increasing levels of software complexity; increasing speed-of-business with hyper-connected users; increasing user expectations for faster apps (more service at lower cost); as well as increasing Lines-of-Business (LoB) demands for faster and cheaper app development and deployment.

It’s certainly increasing the apps performance pressure in the corporate world. While 2-3 updates a year might have been the norm ten years ago, today it’s more like every month. In organizations that have adopted the DevOps working mode, blending the application development cycle with ongoing business processes, it can be a lot more frequent. Flickr does it ten times a day, and Amazon every 11.6 seconds! DevOps evangelist Gene Kim in the Phoenix Project claims that DevOps companies deploy code thirty times faster with 50% fewer errors, and are twice as likely to attain app profitability compared to their non-DevOps competitors.

APM necessary to keep pace with demand

In order to manage the accelerated apps maintenance and update schedules, companies need a single system across the whole lifecycle rather than multiple tools for managing, developing, and supporting usage. This is the task of APM (Apps Performance Managment) software, in this rapidly expanding market area, where all the major IT vendors are active. Typical benefits gained from solid APM efforts include faster customer check-ins and check-outs, better scalability, more efficient use of hardware resources, and, most importantly, better understanding of the real user experience through capture of user interactions with the app — identifying where usage diverges from the expected usage pattern and where customers drop off unexpectedly.

However, continuous measurement of the end-user experience requires data to be generated across the WAN, nationally, regionally, and in some case worldwide. This is why a leading APM provider like Compuware operates a global 100,000 plus synthetic agent network generating test traffic for its customers alongside the customers’ own measurements of their user traffic. This helps customers and Compuware APM conduct instant triage both on hardware and software as well as networks to ascertain where problems are located. Such root-cause analysis reduces problem resolution from days and weeks to hours and minutes. It also enables pre-testing of new apps and updates in WAN environments.

Monitoring these activities in real-time also allows IT to align apps performance measurements with corporate business metrics like conversion rates, bounce rates etc. It also allows leading edge companies to capture KPIs (key performance indicators) in the test phase and continue into production and deployment modes.

Going a step further and expanding APM to also look at the underlying infrastructure allows DevOps teams to look at both platform and app performance. Ideally, this promotes continuous integration with rapid feedback and automated testing, thus uniting pre-production, development and deployment in the DevOps process.

SI’s are gunning for the APM market

We are seeing system integrators like Accenture and Wipro investing heavily in APM (and cloud APMaaS services) and aiming to take away business from corporate IT, by demonstrating to the LoB (lines of business) that they don’t have to wait for their own IT department to catch up.

So, how far is your company along the app development continuum: waterfall – agile – lean – continuous integration -continuous delivery – continuous deployment – continuous operations? European and US data indicates that the majority of large companies are somewhere between ‘agile’ and ‘lean’ today, whereas leading edge companies like Spotify are organized in cross-functional build and run teams for continuous deployment.

APM supports the shift towards more LoB-IT convergence, breaking down corporate structures along the way — concurrent development/deployment is needed to keep up with increased speed of doing business today and tomorrow.