Few doubt the increasing popularity of mobile devices, specifically smartphones and tablets. Their numbers grow by the billion, with 2 billion already in use and, if current trends hold, likely to exceed 5 billion by the end of the decade. While not all of these will connect to an enterprise IT system most will. Increasingly users are consumers of a wide range of enterprise products and services, whether these be in banking, travel, insurance, online shopping or whatever. Such interactions link at some or many points to traditional backend systems.
One challenge for IT is quantifying the impact of so many mobile devices using existing IT and infrastructure. Most enterprises have carefully planned for specific workloads, often sourced from clearly understood communities of users who might be employees, agents, partners, suppliers or whomever. Even web sites which host internet-sourced interactions, such as for searching for information or buying, have largely been designed for specifically anticipated browser session levels, with some overage capacity available to be brought online if needed to support demand.
The arrival, in a relatively short period, of so many smartphones and tablets has the potential to upset IT’s carefully designed apple cart, unless IT is prepared. Already one UK bank processes over 50B transactions a year and Visa acknowledges processing in excess of 45K transactions a second at peak loads. In both cases transaction are increasingly being initiated by mobile devices and are driving significant volumes increases.
One source of change is that many smartphone and/or tablet users no longer feel limited to ‘browsing’ for information like account balances or claims status or airline arrivals and departures. More and more ‘transactional activity’ is occurring because smartphones and tablets are close to their owners so much more of the time than their desktop or laptop. In addition many apps ‘ping’ servers without users consciously needing to do anything. Whatever the specific reason this produces increased levels of activity involving backend systems, those which IT are supporting, or there is the risk of losing business.
To be prepared for this increasingly mobile environment IT needs to know what is happening with its workloads and, if they are changing, why. The key questions that should be asked are how much:
- Additional traffic attributable to mobile devices does the infrastructure support today?
- What volumes should be anticipated in the future?
- Do you know which applications are subject to unpredictable peaking?
- How you will manage those peaks when they occur?
These are not simple questions to answer. They require accurate data. Few applications initiated on mobiles are currently coded in ways where IT can readily separate out traditional and mobile interactions and transactions. To make matters more still complex, many mobile device interactions may start yet may not complete, either because of mobile comms difficulties (increasingly rare) or because of some immediate user inconvenience or even because existing application, or browser, sessions are still not fully adapted for an increasingly mobile environment. (Some suggest that over 50% of interactions fail because users ‘give up’ if there is too much of a delay). But for IT failed logical transactions represent wasted resources which present both application design as well as operational issues for resolution.
This makes identifying which proportion of workloads are attributable to mobile interactions less than straightforward. One illustrative example of such difficulties comes from a Spanish bank that, when asked, replied that it did not know what effect mobile devices had on its IT and infrastructure.
If the most complex of organisations, like banks, have difficulty it is likely that other enterprises will have similar or greater challenges working out the processing impact of mobile devices. For this all need sophisticated tracking tools and even those available may not be sufficient. This applies irrespective of platform, with the possible exception of cloud-based backend processing delivered by the likes of AWS or Azure where the monitoring tools implemented may possess the necessary sophistication due to the nature of cloud-based computing.
Understanding what IT resources are being consumed and why is becoming ever more important. IT needs process consumption baselines so that comparisons can be made. If significant differences emerge these can then be investigated – not least for whether mobile technologies are a cause. The impact of mobility is an issue that many CIOs will face sooner or later, as traditional IT resources that were expected to be sufficient for years incur unforeseen loads and need upgrades.
For the longer term what seems apparent is that consistent identification of mobile-initiated processing is a necessity, both for negotiating with vendors and for monitoring what is actually occurring. Undertaking this early IT can be prepared and then be equipped to drive appropriate bargains.
Content Contributors: Charles Brett