Blog

Managing capacity for seasonal peaks

The case: Posten Norge

---

Background

Posten Norge AS is a Nordic postal and logistic group that develops and delivers integrated solutions in postal services, communications, and logistics, with the Nordic region as its home market. The company is owned by the Norwegian Ministry of Trade, Industry and Fisheries.

The business is quite stable throughout the year but with a high seasonal peak starting with Black Friday and terminating right before Christmas. During this period, the number of “events”, defined as the number of times where a parcel is scanned, increases significantly creating an extra strain on both servers and Mainframe utilization.

Being able to precisely forecast the IT capacity needs and identify bottlenecks is important for two reasons. Too little capacity means the delivery trucks end up waiting on the IT systems – resulting in wasted time and dissatisfied customers and employees. Too much capacity can result in high and unanticipated costs. Either way the result can be extra expenses for the overall operations, lower customer satisfaction and ultimately, missed business opportunities.

---

The challenge

In the coordinating with its Mainframe outsourcer, Posten Norge needs to make sure to reserve the amount of mainframe and server capacity needed as precisely as possible. Too high capacity reservation would be a waste of money and resources. Too low capacity reservation would result in waiting time, lower customer satisfaction and ultimately last-minute adjustments with consequent extra invoices from the outsourcer.

Posten Norge’s environment is quite complex and composed of both a mainframe environment and a distributed environment with many hundreds of servers also in a virtualized setup. Having a total and transparent overview of this environment is key to being in control and securing predictability.

---

The solution

To secure smooth operations at the lowest possible costs, Posten Norge has teamed up with SMT Data, holding quarterly meetings before, during and after the Peak. This ensures preparation, supervision, technical assessment, and evaluation of the situation.

Together with SMT Data, Posten Norge has systematically worked to improve its season peak readiness. As a first step, utilizing the ITBI solution, the company has mapped the capacity utilization to the business objectives and calculated the total capacity usage per parcel. At the same time, an evaluation has been made about the expectations of business volume growth compared to the previous year. By knowing the total capacity usage per parcel and the expected amount of transactions, it was possible to create a good estimate of the expected capacity need.

To lower the baseline of the capacity needs per parcel, together with SMT Data, Posten Norge has identified bottlenecks and inefficiencies from the infrastructure – both Mainframe and servers. These programs have been analyzed and optimized in terms of their impact on the IT performance, capacity and cost.

Finally, in the server environment, critical servers have been identified based on the previous year’s utilization. Agents have been created to automatically monitor the critical servers. These agents send warnings to the IT Operations Department if the capacity utilization surpasses predetermined levels.

---

The ITBI^TM Advantage

Posten’s cloud hosted ITBIaaS solution, collects data from the company’s entire Mainframe and server infrastructures, leaving Posten Norge with a total overview of its capacity utilization across these platforms.

Having ITBI installed for a longer period allows for historical data analysis not only on the Mainframe but also on Server Environments allowing for the development of enterprise capacity strategies, -budgeting and outsourcer negotiations.

“Having ITBI installed on our infrastructure allows us to have the total overview of our capacity utilization. ITBI allows us to perform analysis for ad hoc projects such as our preparation for the Black Friday and Christmas peaks in a cost- and time-effective way thanks to the granular historic data collected from both our servers and our Mainframe installation.”