Uptime´s long-time customer T grupp had a problem. The traffic on their online long distance ticket sales system Tpilet fluctuated heavily being 3 times the amount on Friday´s and Saturday´s peak times than on any other weekday.
They needed a custom solution that would enable automatic up and down performance scaling of the system.
Legacy software that was running on in-house servers where automatic scaling is not a default feature, had to be moved into cloud. Software had to be transformed to fit into Microsoft Azure to enable auto scaling and updating. Current solution had to be repackaged/developed to Azure building blocks like web API, database service, API Apps etc.
We rewrote the software, migrated it into Azure cloud, while the system was live and running 24/7, and updated web server with proxy configuration to support big loads. At the time of switching to cloud, the town time was counted in minutes not hours. Now we have a hybrid system that runs in cloud, leveraging all the cloud benefits. At the same time, we have synchronized the second site on premises for disaster recovery situations.
Group CTO Raimo Seero introduced this project at the recent Cloud Conference, held by the business media enterprise Äripäev. Raimo was one of the keynote speakers, invited to share his insights of what to consider when it comes to migrating legacy software into cloud – moving business critical systems that are in daily use.
Raimo explained why migration of business-critical software should be trusted only to experienced IT partner: “Moving into cloud, you need to pay more attention to recovery plans than you do for in-house systems. Since you cannot control reliability of an external infrastructure, the software has to be durable and dependable – if a cloud service support fails, the software must be able to manage on its own.” Also consider that in cloud, data might not be managed and backed up the way you presume or are used to.