Resulta: Driving global expansion with Kubernetes and Rancher Prime

Destaques

  • 80% faster to spin up a new environment — from hours to minutes.
  • 80% faster to deploy a new region — from days to hours.
  • 80% faster to resolve issues — average 20 min resolution time.

Produtos

Resulta is a global marketing services pro­vider specializing in affiliate marketing and publishing. The company, which describes itself as, “Proudly Nova Scotian and fiercely world-class,” resides in Halifax where it em­ploys 150 full-time staff comprised of expert brand and campaign managers, copywrit­ers, developers and coordinators.

These in-house staff are supported by more than 150 contractors around the world, who help Resulta deliver on its mis­sion to solve complex marketing problems. For more than 25 years, the company has used its dynamic approach to help brands develop web products and digital strate­gies, optimize their performance and make the most of business intelligence in their creative solutions.

At-a-Glance

After a period of rapid growth, Resulta aimed to expand its operations and services overseas. To do this it needed greater efficiency and agility from its technology infrastructure than server-based architectures could provide. Ku­bernetes stood out as the right option. After deploying Rancher Prime, Resulta gained the efficiency, agility and scal­ability it needed to hasten its interna­tional growth.

The journey to Kubernetes

Global expansion is at the heart of Resulta’s mission, hardly surprising given the major growth the company has experienced over the last few years. In a short space of time, it has grown from 24 to 150 full-time em­ployees. Now, the IT operations team, led by Jamie Fifield, Kevin Bruce and Ilya Levin, manages web properties all over the world.

The company, which started life in 1993, kickstarted its global initiative in 2016. Ini­tially focusing on clients based in North America, the team began their internation­al rapid expansion, taking on new custom­ers south of the border, and branching east out across the Atlantic Ocean all the way to the Pacific. Resulta is now priming its tech­nology stack for further diversification.

With a technology-first ethos, infrastructure evolution has long been a key component of Resulta’s expansion strategy. It was— and still is—using a LAMP (Linux Apache MySQL PHP) Stack in Amazon; a set of open source software that can be used to create websites and web applications, alongside virtual machines.

Given the somewhat primitive nature of the stack, Fifield and his team found it chal­lenging to expand into new territories. In­stalling a new server in India, for example, was no trivial task—requiring dedicated re­sources and serious development time. Like many companies, the IT team viewed cloud migration as the natural next step toward increasing agility and removing some of the heavy-lifting associated with having a server-based architecture.

The team’s cloud strategy soon acceler­ated. Embracing AWS gave immediate ac­cess to data centers in multiple different regions. The company then moved toward containerization as a way to scale interna­tionally with limited resources. The team started with Docker. Soon, though, the limi­tations of the platform prompted second thought. With momentum building around Kubernetes, and Docker’s “walled garden” approach losing favor, Resulta decided Ku­bernetes was the right option.

“With a focus on building our international capabilities, it’s critical that our architecture can scale at speed. Kubernetes and Rancher Prime bring an agile approach to IT that is making our vision a reality.”

Why Rancher Prime?

The decision then became which container management platform to choose. This was a simple decision for the team to make as they had prior experience working with Rancher Prime. Given fondness for the platform’s flexible, heterogeneous nature and active community, the decision to work with Rancher Prime required little discussion.

The rationale for bringing a platform like Rancher Prime on board was always to em­power teams of developers to own more of the end-to-end software development and deployment process. Fifield and team envisioned a simplified, streamlined man­agement experience that would enable developers to oversee projects from build, to test, through to deployment without the need to involve operations. Why? The more autonomy and power given to developers directly, the fewer the project overheads and delays.

For this to work, the company has added a layer they call the “GitOps Model,” which enables developers to control the entire workflow. Having previously worked with Docker Swarm, the team already had ex­perience with containers, so the switch was natural. The team now simply branches, tags and goes to a specific URL to access projects.

What are the challenges Resulta has overcome?

International expansion

Since 2016, Resulta has been keen to ex­pand its client base globally, so it needed a platform that would enable it to spin up a new environment in each country in a seamless way. By running its infrastructure in Kubernetes and Rancher Prime, the team found scaling a pain-free process. Once Resulta had its Rancher Prime environment in place, spinning up a new instance was as simple as “popping up a tent.” What’s more, now that the team has begun the process of migrating specific brands to the plat­form, they are confident about managing a greater number of clients with the same number of staff.

Developer empowerment and ease of use

Empowering its growing cohort of develop­ers was instrumental in prompting Resulta’s journey to Kubernetes. The IT team wanted developers to “own” more of the process and so they were keen to equip them with the tools they needed to do their work more efficiently. Working in Rancher Prime, it now takes developers 15 to 20 minutes to spin up a new environment—it’s simply a case of adding new instances from different re­gions into existing clusters. Regional expan­sion takes longer, but Resulta describes the process as “easy.” It now takes just hours, rather than days to deploy in a new geog­raphy.

Equally, it’s been a pleasant journey for Fi­field and his team who now enjoy greater visibility due to the fact they can see all their clusters and workloads through a single in­terface. Rancher Prime enables them to see their sites being deployed automatically, which the company describes as a real “Aha!” moment. The team currently has the majority—around 90%—of back-end servic­es running in production in Kubernetes and Rancher Prime.

Quality assurance and on-demand support

Reducing support overheads was anoth­er pain point Resulta was keen to over­come. Previously, production issues could take hours, if not days to resolve. With Rancher Prime, the process is simplified. After fil­ing a ticket, teams typically receive a re­sponse within 20 minutes, after which any issues are promptly resolved over Zoom. The company has also been making use of Rancher Prime’s 12,000-strong online community, where they have been able to get rapid responses to pressing technical questions.

This improved efficiency is particularly im­portant when it comes to Resulta’s quality assurance (QA) process. Previously, the QA team would spend valuable time waiting for a member of the operations or DevOps teams to push a project along, which led to wasted time and delays. With Rancher Prime, the team can quickly check code and add a tag to a branch—these efficiencies are a boon for the QA team.

Reducing cloud overheads

As a company all-in on AWS, the compa­ny’s cloud costs are a growing concern. Resulta is now actively looking at a multi-cloud strategy to mitigate future rising costs. This is a decision that has been fur­ther spurred by availability issues in certain regions. AWS currently has no data center in South Africa, for example, and its Brazil­ian centers often struggle with reliability is­sues. With that in mind, and a desire to re­duce increasing overheads, the company is keen to work with the Rancher Prime Ku­bernetes management platform to more closely analyze its cloud costs and to make the switch to multi-cloud as easy as pos­sible.

What’s next for Resulta?

 

The next 12 months for Resulta will see the company continue to focus on global ex­pansion. While around 90% of the compa­ny’s back-end services are running in Ku­bernetes and Rancher Prime, the IT team is currently in the process of migrating their clients’ WordPress sites to the platform; it has 29 sites in production right now, with many more to come.

The year ahead will also see Resulta focus on continuing to create efficiencies across its evolving architecture. By giving devel­opers the tools they need to do their work more efficiently, the operations team can step back, transition accountability and fo­cus on taking Resulta to the next level.