You're Never Too Big to Think Like a Startup: Disrupting the Market with Cisco
In today's market, companies need to move fast. Large and established businesses have to pivot with little notice to compete against lean, agile startups. As a large company, the best way to disrupt your industry is with the latest technology and innovative ideas.
A couple of years ago, T-Systems had the opportunity to make an even larger impact in our market. A subsidiary of Deutsche Telekom, T-Systems is an end-to-end IT service provider with more than 30 data centers spread across the globe. Based in Frankfurt, Germany, our German and international clients include many leading multi-national organizations and we employ more than 38,000 people around the world.
With so many resources spread far and wide, our network infrastructure is massive. We had to find a way to further simplify something that was, by its very nature, complex.
Thousands of Endpoints and Complex Systems
T-Systems operates more than 100,000 virtual machines (VMs) worldwide at any time, and our number one business in MIS&PC these days is helping companies move their operations to the cloud. We are primarily known for our client-centric culture and our expertise with traditional data center architecture, but we have also built a reputation for managing state-of-the-art technologies, like hyperscaler services and software-defined infrastructure.
Part of our work is to provide connectivity between our clients' physical and virtual machines—regardless of where they reside. We work with multiple clouds, private and public cloud flavors, dozens or sometimes hundreds of applications, and thousands of endpoints across the world. Our engineers have architected bespoke infrastructure solutions alongside cloud solutions that provide security, capacity, and fast connections across vast distances. These network designs must meet our customers' business needs, and also comply with local and international regulatory frameworks.
To ensure our customers' systems remained visible and easy to manage, we had to streamline our own architecture and create efficient network administration workflows. But this can be complex with traditional hardware-based server and network architectures. It became clear that to most effectively support our customers, we needed to focus on IT transformation by moving from a hardware- to software-centric approach leveraging latest software-defined technologies.
Following Our Customers' Lead
As an IT service provider, our first step in a relationship is always to understand how our customers do business. We may be able to recommend a newly released network device or a proven solution that's worked for us in the past, but it may not fit their particular needs. After all, we look at how we’re best suited to help them improving their business outcomes. We also help our customers anticipate their future needs. But sometimes it is hard to predicts what’s coming; especially in disrupting times we are in. In such cases we must be able to react fast and in an agile manner to deliver against customer demands.
Over the past years, it had become apparent that we needed to enhance our existing network infrastructure and our operations model to continue meeting our customers' expectations. Among our top challenges were maintaining or reducing service costs, improving uptime and stability, reducing resolution times, and rolling out new infrastructure for our clients much faster. On top, we needed to find a more efficient way to problem-solve and troubleshoot, which involves dealing with several hundred of changes per day. And we knew the only way to keep up with such demand was to continue following our customer centricity model and embrace automation to an extend we had not seen before.
We had three criteria in mind for a comprehensive solution. The first was end-to-end automation, including zero-touch provisioning and remote management from a single pane of glass. The second was a carrier-grade Telco platform that could deal with highly distributed environments.
The third and most important consideration was our desire to become independent of hardware. We wanted to shift our focus to software-defined infrastructure that would simplify management and allow our customers to better compete in today's fast-moving business environment.
Searching for the Perfect Fit
We started our search in earnest with an RFP. In addition to our three criteria, we also established a thorough matrix to evaluate the technical and commercial potential of our vendors' proposed solutions. The entire process, including the RFP, proofs of concepts, tests, market evaluations, and contract negotiations, took about eight months. At the end of this selection period, Cisco won the bid.
T-Systems has a long-established relationship with Cisco. The company has been one of our primary technology partners for decades, but we had to perform our due diligence. Fortunately, on paper and in practice, the company presented the solution that best fit our needs.
From POC to Production
The transition to our new architecture began the moment we signed the contract back in October 2019. We started by converting our proof of concept into a development environment at our Frankfurt data center and testing 12 use cases that represented our customers' most pressing needs. In April of 2020, after months of exhaustive tests, we created a mirror of our development platform at this same data center and rolled it out as our first production environment using our new Cisco tools and the first Multi Cloud Connectivity Platform (MCCP) was online.
This new platform helps our customers solve the hybrid cloud connectivity challenge as it allows for easy integration of public and private cloud capacity into their infrastructure. The days of waiting for switches and servers to ship are over. Our customers can fast-track new connections and capacities without the long-lasting planning and provisioning cycles they were used to in the recent past. Now, they can easily build complex networks spanning from on-prem to T-Systems data centers, all the way to the public cloud with Google, Microsoft Azure, or AWS. The foundation of MCCP is Cisco ACI Anywhere, which we chose because we can deploy it at every point on our global network.
A few weeks later, we rolled MCCP out at our Magdeburg data center, and we are now deploying it at our North American data center in Houston, Texas. The transition for our facilities in Brazil and Singapore is next on our agenda. By the end of 2021, we hope to have completed the transition of all our strategic data centers to software-defined infrastructure.
Orchestrating a Complex Network with Ease
Our new network provides the same complex functionalities as our old one, and more. Yet, it is easier to operate. MCCP allows us to programmatically manage, maintain, provision, and troubleshoot our infrastructure from a central location with support of Cisco Network Services Orchestrator (NSO). We can set up all our data centers from one central operations center instead of relying on separate teams at each of our international locations. As a result, we are both lowering costs and ensuring that all of these data centers are on the same page.
We can now give our clients the ability to ramp-up their infrastructure using a simple visual interface. Instead of asking our engineers to spend time architecting a solution, I can send a T-Systems customer to our online portal to add virtual machines, increase bandwidth, and expand their network at the press of a button. They can generate a request in minutes, and our platform can then automatically deploy their new infrastructure within hours.
When we decided to make changes to better serve the future needs of our customers, we rely on our experience along with the support of Cisco technology in order to make these changes a reality. These tools didn't even exist a few years ago, and now, we can ensure our infrastructure is operating optimally at all times and help our customers swiftly navigate today's increasingly competitive business climate.
Agility and Adaptability
New technology creates new opportunities. For example, disaster recovery was one of the 12 use cases we set out to explore during the infrastructure development phase that followed our initial proof of concept. Now, we can now offer these services to our customers in partnership with Cisco.
Cisco has provided valuable assistance in the form of training material that has helped us eliminate silos, master scrum methodologies, and accelerate our release cycles. A component of our partnership with Cisco is continuous improvement and evolution of the solution. Together, our joint team creates new or updates existing features every two weeks, which we then incorporate into quarterly platform updates, so our customers are consistently benefiting from our improvements.
Embracing the Future Faster
We’ve made so much progress over the last few months, it’s astounding. Careful planning and minimizing risk is important for a company as big as ours, but we also have to be brave. Moving to software-based infrastructure is a big step, but it’s now put us in the best position possible for the future.
As a large company, T-Systems can be agile and flexible—this is part of our mindset and DNA. We have adopted agile and lean methodologies and are operating entities like our infrastructure DevOps team in startup mode.
Everyone wins when companies constantly strive to improve, and choosing a reliable partner is critical for these journeys. Here at T-Systems, we never stop innovating, and we know that Cisco will support us on our innovation journey today, tomorrow, and for years to come.