Subscribe to get the latest
#82 Shifting from Infrastructure to Workloads
on Wed Mar 30 2022 17:00:00 GMT-0700 (Pacific Daylight Time)
Darren Pulsipher, Chief Solution Architect, Intel, continues his conversation with Sarah Musick, Cloud Solution Architect, Intel, about the shift from infrastructure to workloads. Please place in the Embracing Digital Transformation playlist.
Sarah’s conversations with customers have shifted from talking about infrastructure to talking about workloads. One obvious reason for this has been the shift from the data center to commoditizing resources to consuming anything as a service. Another reason is the influence of app teams in architectural decisions. The applications are the lifeblood of organizations in a way that they were not even five years ago. The last reason is financial: budgets are now categorized by workloads as opposed generalized IT spend.
One challenge in this shift from infrastructure to workload is striking a balance between centralized and decentralized processes and workloads. Some things, such as security governance, are handled by a centralized hub, and others are handled in local, smaller teams. Much of the decision-making is being pushed down into the teams that are creating new applications and services inside IT and also just for customers.
COVID was an accelerant to promote change and move organizations to the cloud. Some IT organizations are letting their applications move to the cloud without constraints, and would rather clean up on the back end than slow down the innovation that is occurring. This “Black Swan” event is unprecedented and we are still seeing the fallout from the quick paradigm shift.
The role of the CIO is back, but only if they start thinking about information and workloads again instead of running a data center. Of course, it depends on the organization. Cloud-native organizations, or those that aspire to be cloud-native, are refactoring their applications in flight because they want to be super agile. The more they do that, the more it drives toward infrastructure and service of the application rather than accepting limits that existed in the data center before and working within those parameters. Previously, in that situation, it triggered innovation from app teams because when you are dealing with a set of givens, sometimes necessity is the mother of invention, as opposed to the infinite possibilities in the cloud
That’s one end of the spectrum. On the other end are American heritage corporations, the institutions. Typically, they still have information on the mainframe. It’s an “if it ain’t broke, don’t fix it” situation, especially with static applications. The cloud is enabling CIOs to think beyond the old school way of managing these applications. If CIOs can embrace these new technologies they can now see a path forward.
The processing that is happening under discrete applications is more relevant than ever. Intel has a huge role in addressing concerns about performance or cost in cloud offerings partly because they have mothered those technologies and also because they are one of the biggest software companies. There is a vast amount of internal knowledge. In other words, not all instances in the cloud are created equal, so Intel has optimized workloads internally to get the most out of the cloud instances they are using.
For example, many organizations are moving things to Kubernetes clusters and Intel does a ton of optimization around that. They can take things beyond the standard Helm charts with extensions that would look at the health of the node underneath and not just raw availability. There are many things Intel can do to help customers vastly improve performance and cost, not just 2 or 3 percent, but 30 or 40 percent?
Every workload does not belong in the cloud. The structure of an organization has an impact on where the workload should land. The key is to be cloud-smart.
A successful multi-cloud strategy is having a primary and a secondary cloud. When many talk about multi-cloud, the motivation behind that is fear of vendor lock-in. Where most of your data lives has a bearing on multi-cloud strategy, as does where workloads fit best.
To develop a strategy, Intel’s cloud solution architects will engage in discovery about what the organization wants to do and where the problems are. Intel often can remediate many of the issues with tools they have at their fingertips. The cloud solution architects will also contextualize offerings to make the process faster and more efficient. Part of their job is to be an educator, so everyone has the information they need to move forward.