DevOps Summit: DevOps flexibility lets Hailo manage peak time and tube strike pressures
Boyan Dimitrov, platform automation lead at Hailo, explains how DevOps has made the firm's coders more productive
Harnessing the power of DevOps has allowed taxi ordering service Hailo to bring about a more iterative approach to application development, which has in turn led to faster execution and an improved service.
That's what Boyan Dimitrov, platform automation lead at Hailo, told the audience during his presentation at Computing's DevOps Summit 2015.
Hailo Hub is an application for hotels, restaurants and other businesses that require taxis for their customers, and Dimitrov described how the business team was able to use fluid techniques to develop it all from scratch "totally on their own".
That, Dimitrov explained, wouldn't have been possible prior to Hailo adopting a flexible development model through DevOps
"Just to give you an example; if we'd had to do the same thing before, working how we were, it would have taken months and probably four times the people. And it would've been out of date within a month," he said.
Now everything has changed and through an iterative approach for fast execution, Hailo is able to work on hundreds of development projects at the same time.
That improvement in productivity has come after moving from legacy architecture onto a cloud-based platform - Amazon Web Services - which has "brought about 30 per cent operational savings", said Dimitrov.
"We can now be more flexible... we scale up as we get more traffic, because as you can imagine, the taxi business is quite peak driven," he said.
"For example, today there is a Tube strike and this means for us our system will scale and then we can easily scale back," he said.
Ultimately, switching to a DevOps model has allowed Hailo to become more efficient and more productive.
"Nowadays we do dozens of daily deploys, either deploying new services or modifying existing services and this is quite different compared to where we were before with bi-weekly deploys," said Dimitrov.
"Now we deliver features in parallel, and if the business environment changes we can quickly align our development," he concluded.