Subscribe to get the latest
#44 Cloud in the Digital Age
on Mon Mar 22 2021 17:00:00 GMT-0700 (Pacific Daylight Time)
In part two of the interview, Darren Pulsipher, Chief Solution Architect, Intel, and Doug Bourgeois, Managing Director, GPS Cloud Strategy Leader, Deloitte, continue their discussion about the cloud migration.
Over the last five or six years, the methodologies, tools, and experience in cloud migration have matured to repeatable processes.
The first step is to decide on your migration priorities because it’s going to happen in phases, not in one big move. This does not take a lot of time or resources, but it is critically important. An extreme example is that you would not want to pick the mainframe as the first system to move to the cloud, rather a more self-contained system such as email.
Companies such as Deloitte have developed and invested in discovery tools that help accelerate the migration processes. These tools will grab a broad set of data, run an algorithm that looks at complexity, and rank all of the systems into different categories. Understanding the configuration and integration points of existing systems and compatibility of software components is fundamental to cloud migration. We also need to look at boundaries and compliance frameworks such as PCI or HIPAA. Building out landing zones for these environments in the cloud is phase two in the process.
At times there is a substantial amount of prep work for migrations. The first wave is the easiest with the least amount of modifications, but after that, in phase two, there may be upgrades or changes to operating systems, re-platforming, or moving to a different types of database, for example. The third wave often involves more antiquated client servers or proprietary architectures that require significant re-architecture and may take months to prepare for cloud readiness.
It’s important to distinguish that cloud readiness is not the same as cloud optimization; that comes later.
In many cases, the driving force behind moving to the cloud is business driven rather than technology driven. For example, a client may not want to continue a lease just to house a data center, or are shifting their physical offices. In those cases, there is a time factor where it makes sense to execute the migration based on readiness rather than optimization.
Once in the cloud, you need to optimize because the cost drivers are different in the cloud than in the legacy data center. The cost of a data center, after making the initial investment, is relatively hidden, whereas the cloud is more of a rental agreement that goes on in perpetuity. Many times in the legacy systems, we solve problems by throwing on more memory, more CPU, or more storage because it works to a certain extent, but this creates inefficient systems. If we simply move these inefficient, resource-intensive systems to the cloud, the cost model is much higher than it needs to be, hence the need for optimization.
Some of the optimization process might be a process change. For example, for an organization in Canada, their cost went through the roof when they did a lift and shift of an SAP instance into the cloud. They realized that they weren’t using this instance at night or on weekends, so they went from a 24/7 model to a 16/5 model. Making this switch saved them a substantial amount of money. So there are ways of making a small effort, high value return with different approaches.
We are finally seeing, after being more than a decade into cloud, an emerging trend of finding value in a change of business strategy rather than in infrastructure. The COVID-19 pandemic was certainly a factor in accelerating this change. A perfect example of this is telemedicine. It already existed, but had been stagnant for five or six years before the pandemic; now this model is the norm.
Transformative innovations are happening in the cloud. As more systems move to the cloud, industries will continue to try and adopt different models with new, transformative capabilities.