<<

Whitepaper The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication Charles Araujo Principal Analyst, Intellyx

April 2020

Sitting in on a tech company sales pitch these days can seem a little bit like a scene from My Big Fat Greek Wedding — no matter what your ailment, Cloud (instead of Windex) is the answer!

In fairness, there's no question that the cloud is already a significant part of the enterprise technology stack and will play an even more substantial role in the future of any enterprise IT function.

Still, the pervasiveness of cloud adoption does not come without a healthy dose of irony.

While much of the pitch about anything cloud-related focuses on its inherent modernity, and with it, the implication that it will solve IT’s dysfunctional woes of the past, cloud management practices are increasingly replicating those very dysfunctions.

Far from realizing a holistic management purview of their newly cloudified technology assets, enterprises are, instead, creating a whole new host of management silos. Cloud performance monitoring, one silo. Cloud cost and bill analysis — another.

And capacity and workload management? Most organizations aren’t even scratching the surface of that one. 3

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

When cloud initiatives were in their early let’s-test-this-and-see-what-

happens modes of operating, this situation was all manageable

enough. All it required was a little elbow grease.

That’s no longer the situation.

If enterprise leaders do not want to replicate the dysfunctions of the

past, they are going to need to take a fresh look at cloud management

— and make sure they get these three things right.

If enterprise leaders do not want to replicate the dysfunctions of the past, they are going to need to take a fresh look at cloud management — and make sure they get these three things right.

©2020 Intellyx LLC. https://intellyx.com 4

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

Right #1: Rightsizing

When it comes to getting cloud management right, the best place to start, as the saying goes, is at the beginning.

A vital component of cloud management is the migration of workloads from traditional, on-premises environments to the cloud. And, this is the first place that many organizations start to go off the rails.

At the very beginning of the shift toward cloud architectures, there was often very little analysis — if any. In part, this was because many initial cloud workloads were for so-called greenfield applications. They were brand spanking new, so there was little analysis to be done.

As organizations began moving production workloads to the cloud, however, complications almost immediately arose. Operations leaders responded by doing some superficial analysis, but as cloud deployments mature and grow, that's proving to be insufficient.

While the first cut at analysis was adequate to keep cloud migrations from crashing and burning, the lack of in-depth analysis often served to bury the problem, making it that much harder to surface later on.

©2020 Intellyx LLC. https://intellyx.com 5

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

The first thing that organizations need to get right if they are to avoid cloud management dysfunction, therefore, is workload rightsizing — and they need to get it right at the point of migration.

To do so, you must dig deeper and analyze the full-stack resource utilization of your target workloads. This analysis should include CPU, memory, I/O, network, and storage — and with enough depth and granularity that it will enable you to do comprehensive what-if analysis as you assess your deployment options.

Most importantly, you need to adopt a workload-centric perspective — rather than a systems-centric perspective — as you employ this analysis.

Not all workloads are created equal. This fact is particularly important as you attempt to strike the right balance between performance and cost optimization during cloud migrations. Likewise, as you complete this full-stack analysis and seek to rightsize your target environments accordingly, you may find that not all workloads are, in fact, viable targets for a cloud migration, or that you should prioritize and sequence migrations differently.

The point is that an effective rightsizing capability will allow you to approach cloud migrations with your eyes wide open — a sure way to minimize dysfunction down the road.

Not all workloads are created equal. This fact is particularly important as you attempt to strike the right balance between performance and cost optimization during cloud migrations.

©2020 Intellyx LLC. https://intellyx.com 6

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

Right #2: Operational Performance Management

After you have moved a workload to the cloud, that’s when the real management work begins.

While moving a workload to the cloud comes with many benefits, it also brings a set of challenges. Ironically, many of those challenges are a direct function of the benefits the cloud offers.

The ability to configure cloud environments in countless ways and employing so many different and nuanced configuration options gives the enterprise immense flexibility and agility — but it also makes the job of monitoring and optimizing those deployments a significant challenge.

The ability to configure cloud environments in countless ways and employing so many different and nuanced configuration options gives the enterprise immense flexibility and agility — but it also makes the job of monitoring and optimizing those deployments a significant challenge.

To add fuel to the fire, most of the configuration options exist on each of the various public cloud providers — but each provider has implemented and configured them in different ways.

As most organizations have adopted a multi-cloud strategy, that means that you must stay up-to-date on these countless configuration options across these different platforms.

©2020 Intellyx LLC. https://intellyx.com 7

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

While this complexity is a challenge from a broad management perspective, operations Achieving this type of holistic teams that are responsible for monitoring and operational performance optimizing cloud workloads may feel it most acutely. management demands that you move beyond simple reports and It is imperative, however, that organizations get operational performance right if they are take advantage of advanced to avoid replicating the dysfunctions of the performance management past and, instead, manage their cloud approaches. environments as part of a cohesive hybrid strategy.

Achieving this type of holistic operational performance management demands that you move beyond simple reports and take advantage of advanced performance management approaches. These approaches should allow you to bring together sophisticated detection techniques with the appropriate context to enable you to assess the situation in this dynamic environment more accurately.

For instance, some modern cloud management systems are leveraging advanced anomaly detection techniques and pre- configured alerting policies (based on industry best practices) to identify potential performance problems in cloud environments. They are then adding context to allow operations leaders to readily assess the situation.

©2020 Intellyx LLC. https://intellyx.com 8

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

An example of this contextual analysis is the detection of an increase in CPU cycles. By itself, it may be insignificant, but if an increase in the number of processes waiting for CPU cycles accompanies it, it is much more likely to indicate a performance issue.

Most importantly, organizations must build an operational performance management capability that is optimized for the various configuration options in the cloud, that allows for the cloud's dynamic nature, and which is application- centric in its foundation.

The goal of this performance management capability must be to integrate into the existing operational management model, provide visibility to potentially impacted operational teams, and enable automated scaling (when possible) to maintain workload performance and avoid application disruption.

Most importantly, organizations must build an operational performance management capability that is optimized for the various configuration options in the cloud, that allows for the cloud's dynamic nature, and which is application-centric in its foundation.

©2020 Intellyx LLC. https://intellyx.com 9

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

Right #3: Cost Management

In another case of the benefits of the cloud also leading to its challenges, all of those different configuration options also make cost management a colossal headache.

While those configuration options offer organizations all the flexibility and agility they seek, each option also comes with different cost attributes. This fact makes balancing operational performance and costs a delicate and intricate dance.

Adding to the challenge, this dance is something for which most enterprises are ill- prepared.

On-premises environments have generally represented sunk costs. An organization made a substantial investment in a technology stack. As a result of that tremendous up-front investment, configuration and deployment changes had little to no impact on operational costs. For better or worse, you could optimize for performance without any consideration to cost — because there were none.

With cloud workloads, the converse is true.

While you may be spared the massive up-front investment, you must build a sophisticated cost management capability to avoid making that investment many times over during the life of your deployment.

The cost management challenge, however, is further amplified because organizations must manage their application stack in a holistic, hybrid manner regardless of whether workloads live on-premises or in the cloud. Not doing so will be a sure recipe for dysfunction as teams manage against different parameters.

To avoid this and effectively manage cloud costs, you must build a management capability with detailed, resource-by-resource spending analysis as its foundation. This type of capacity will enable you to drill down into the specific configuration options and attributes that are contributing to costs and allow you to perform what-if analysis as you seek to balance cost and performance across your hybrid environment.

©2020 Intellyx LLC. https://intellyx.com 10

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

Implicit in this capacity, however, is a set of cloud-specific capabilities that are central to your ability to fine-tune cost management. For instance, you must be able to detect both idle resources and over-provisioned ones — and both in the context of current and forecasted capacity demands.

Moreover, you must be able to perform this analysis in a way that results in specific recommendations for configuration tuning that takes into account both workload demands and resource usage in the context of both performance and costs.

Moreover, you must be able to perform this analysis in a way that results in specific recommendations for configuration tuning that takes into account both workload demands and resource usage in the context of both performance and costs.

The Intellyx Take: It All Comes Down to Balance

So there they are, the three things you need to get right to avoid the dysfunctions of the past and build an effective cloud management capability.

I can sum them up like this: start on the right footing, focus on the unique requirements of performance optimization in the cloud, and respect the cloud’s cost management complexity.

But perhaps I can sum it up even more simply: build a capability to continually and dynamically balance cost and performance.

That balance is really what it all comes down to.

©2020 Intellyx LLC. https://intellyx.com 11

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

Striking that balance, of course, is much easier said than done. Achieving it will demand that you make wise choices when it comes to both your partners and your tools.

It will be imperative that you select cloud management tools, such as Virtana’s CloudWisdom, that can help you with all three of these areas. Ideally, your chosen tools will do so in an integrated, holistic fashion and allow you to manage it in a single or unified interface.

Moreover, it will also be critical that the tools you select provide simultaneous visibility into both your cloud and on-premises environments. This capability will be essential as you seek to measure workload performance during migrations and will enable you to manage your hybrid environment in an integrated fashion.

Attempting to manage your on- premises environments separately from your cloud environments will both perpetuate the dysfunctional silos of the past, and be a recipe for failure.

Finally, it will be critical that you understand that cloud management is, by its nature, a dynamic activity that will remain in a continual state of change. It will, therefore, be essential that you adopt a management framework that balances performance with cost, and which enables you to continually adapt your cloud architecture to your ever-changing business needs.

It is only in finding this balance that you will finally be able to realize the full potential of your cloud investments — and the promises of a different future.

©2020 Intellyx LLC. https://intellyx.com 12

The 3 Things You MUST Get Right with Cloud Management to Avoid Dysfunction Replication

About the Author: Charles Araujo

Charles Araujo is an industry analyst, internationally recognized authority on the Digital Enterprise and author of The Quantum Age of IT: Why Everything You Know About IT is About to Change.

As Principal Analyst with Intellyx, he writes, speaks and advises organizations on how to navigate through this time of disruption. He is also the founder of The Institute for Digital Transformation and a sought after keynote speaker.

He is a regular contributor to CIO.com and has been seen in Time, InformationWeek, NetworkWorld, Computerworld, USA Today, and Forbes. About Virtana

Virtana is the leading AIOps platform for digital transformation. Our technology and services give innovative organizations the clarity they need to take control of their infrastructure, transform their cloud operations, and deliver a superior brand performance.

Virtana’s software modernizes IT, supporting its agility while guaranteeing performance, minimizing risk, and reducing cost. We guide users through a journey to see their infrastructure from a single pane of glass, ACT on issues that arise, and transform processes to automate for the future. Follow us for industry insight on Twitter | LinkedIn. Virtana: Take control.

Copyright © Intellyx LLC. As of the time of writing, Virtana is an Intellyx customer. Intellyx retains final editorial control of this paper. Image credits, in order of presentation: Claudio Schwarz, Siora Photography, Christophe Hautier, Noel Nichols, chuttersnap, Jungwoo Hong, and Drew Beamer.

©2020 Intellyx LLC. https://intellyx.com