Next Tech Insider: Cloud Migration - mistakes made and lessons learned
Next Tech Insider takes a look at transformational technology from the perspective of leaders within the industry. In July Michael Borts, Senior EVP, Head of Digital Transformation joined the Cloud Talent team at Empiric to share his experience of cloud migration projects in the banking sector. He reflects on mistakes made and lessons learned.
“Very few bands that start in a garage actually turn into deals. Most garage bands just die unknown.”
If Michael has learned anything in his 20+ years as a technology leader, it is that when projects are poorly executed, it is usually because a group of people get together and says, ''let's just do cloud''. It rarely ends well without expert guidance.
“I'm a former architect” explains Michael, “a real brick and mortar architect, so if you want to compare Cloud Migration to building a structure then consider that when people build their houses, they need to construct a frame over a sound foundation, build in the wires and plumbing, and the network lines. If you don't have these key features executed properly, what you build on top will not function”.
Michael started his technology career at Merrill Lynch, where he worked on the world’s first billion-dollar project in the IT industry. Named the ‘Trusted Global Advisor’, and marking a complete overhaul of the company’s IT systems, the project was essentially a client server system, a blend of off-the-shelf software and customised applications to provide state-of-the-art financial analysis and customer service software to some 25,000 brokers and admin staff in 600 officers across the United States.
After a stint at Bear Stearns, he moved to the New York Stock Exchange which he describes as ‘probably one of the best employers’ he ever had in his life. Everything was thoughtfully designed and delivered; every hire, every project, and every project plan. Nothing went undocumented and nothing was executed in bulk. Each process was tailored for each individual situation and this led to the Exchange becoming renowned for its high-quality technology.
How to properly migrate to the cloud
Currently at one of the industry leaders when it comes to cloud thinking, he is writing their cloud migration program, a program which has taught them a great deal about how to properly migrate to the cloud.
“Most projects are rarely on time,” says Michael. “Software projects tend to run at best 30% to 50% longer than expected, at worst, three times longer. And there are increasing costs with that. Our goal is to control the timeline, reduce the delays, avoid any re-work and provide guidance to every impacted team”.
In cloud migration projects, a new cloud environment is often called a landing zone, the foundation on which a cloud migration project is built.
“If done poorly”, he says, “it most definitely causes downstream delays and cost overruns. Being a garage band for cloud migration is risky.” It is here where professional assistance and guidance are needed. Most successful teams bring in the cloud experts as full time hires or third-party partners. AWS Professional services consultants are one such option for Amazon technology.
Once the landing zone is built, the teams can start moving their workloads to the cloud, application by application and system by system. Here, pervasive automation needs to be deployed, else the project will fail in one of what Michael describes as ‘multiple painful death scenarios’.
“Either I build it and it's so expensive I can’t increase my cloud presence because nothing I did is automated”, he states. “In extreme cases, I have to shut it down. Or I am trying to build it, but I am just replicating my terrestrial environment in the cloud, pouring concrete into the cloud. I can’t support it and my SLAs are broken. In some cases, I have the technology, but I don’t have the expertise to continue to build my cloud out and have to slow down my migration to enable my organisation to catch up.”
Which model to follow?
Another one of the most common challenges with cloud migration is whether to federate or not to federate its ownership and operations. A federated model allows a business to optimise the enterprise IT service delivery. It is agile, fast and it is fully owned by the product development teams. However, it is also an exceptionally costly model compared to centralising some of the services which avoids having to replicate the same skillsets across different areas and teams.
Michael likens this concept to the United States of America without a federal government. It could function but it would be almost prohibitively expensive and is extremely difficult to manage. In his opinion, nobody has been really successful running a true DevOps federated operation. The right model he feels is a combination of some global shared services and local development teams’ ownership, sometimes referred to as a ‘hub and spoke’ model.
Michael points out that not everything was born equal: “If you follow Amazon literature, they'll tell you they use small teams. Okay, what's a small team in Amazon's language? Well, it's a two-pizza team. People who can be fed with two pizzas. That, by the way, in the USA is 8 to16 people.”
Many businesses do not have 8 people development teams. As few as two people can be the norm, hence, the need to have a centralised shared team. However, that itself brings in a whole different set of challenges. Centralised services may shrink headcount in comparison to federated services where headcount increases due to the need to augment every team, but the skills are needed at a different level of expertise and often are more expensive. Michael describes such shared expert teams as the ‘Swiss Army Knife' of cloud development.
Cloud cost management
Building cloud services is becoming more and more routine but it seems that the industry is just beginning to realise the need for proper cloud operations. Such knowledge is still rare.
A large component of this operational skillset is the cost management of a growing cloud environment. “As an example, every scale up of a database server doubles its price”, explains Michael. “If you continue to scale up across the workloads but have no elasticity to scale back, a few months later you'll be spending millions just in development and other non-production cost.”
So, could the next trend in the future of the cloud be cost management and the running of clean environments? And if so, how quickly is the industry moving towards improving cost efficiencies?
“Architecting for cost is important”, agrees Michael. “But right now, I don't think the industry has awakened to that yet. They architect for efficiency and speed to market, and for simplicity of development and compute power. But few teams are architecting for cost. We are not there yet.”
Simply provide us your contact details and we will be in touch
Empiric is a dynamic technology and transformation recruitment agency specialising in data, digital, cloud, security and transformation. We supply technology and change recruitment services to businesses looking for both contract and permanent professionals.
Empiric are committed to changing the gender and diversity imbalance within the technology sector. In addition to Next Tech Girls we proactively target skilled professionals from minority groups which in turn can help you meet your own diversity commitments. Our active investment within the tech community allows us to engage with specific talent pools and deliver a short list of relevant and diverse candidates.
For more information contact
To view our latest job opportunities click here.