This blog is the first in a series on cloud optimization. Join us the next few weeks as we explore why moving applications to the cloud can result in higher than anticipated cost, how to plan ahead to avoid overspending before migrating to the cloud, and best practices in optimizing workloads and cost in multi-cloud environments.
The following is an interview full of useful insights. In this interview, I talk with Adam Lubsen, Cisco worldwide business development manager for workload optimization products, about why IT leaders are finding their public cloud costs can be greater than what they may be able to sustain and what you can do about it.
Q: Remind us why moving applications and workloads from data centers to the public cloud is such an enticing idea.
A: Quite simply, the cloud changed everything. What I mean by that are two things. First, the promise of the cloud is to make it easy to spin up infrastructure to incubate and develop new applications to drive innovation. And second, because cloud providers have ready-made infrastructure, offered through a consumption model, it can be cheaper to both develop applications and run them in the cloud.
Before the cloud, it took time and money to purchase new hardware and software and set up environments for developers to work on new applications. The cloud came along and changed the paradigm. It immediately opened new paths for innovation by removing IT hurdles around capital expenses and infrastructure. Developers got a ready-built playground with quick access to everything they need to try new things. Plus, the cloud is elastic. You can scale up or down based on your requirements. So rather than justify up-front investment and hope you got your requirements right, the cloud lets you move fast and focus on building your business using on-demand, rather than fixed, resources.
Q: Why can costs for public cloud be greater than what IT teams may have anticipated?
A: Moving to the cloud isn’t just flipping a switch. It requires planning and a different mindset. Yet often we see a few things happening. First, IT teams are so focused on not compromising application performance that they over provision cloud resources. They don’t know what resources are needed to ensure app performance in the cloud, so they overspend from the start.
Second, when planning does happen, it usually is offered by a return on investment or total cost of ownership assessment from the cloud provider. The assessment typically includes two views, one for lifting and shifting the application and a second view if you optimize the application prior to migrating it to the cloud.
The issue with these assessments is that they’re tied to a specific point in time. Whether or not you simply move the application as-is or optimize it before moving it to the cloud, these assessments can’t see into the future. Once your application has moved to the cloud, application resource demands will keep changing, the resources required to meet the needs will keep changing, and the services and costs from your cloud provider will keep changing.
But if you have a data center mindset, you won’t be prepared to operate efficiently or cost-effectively in the cloud. For example, you’re accustomed to leaving resources “on” because they were a fixed cost in the data center. But in the cloud, you’ll keep paying for stranded storage volumes and other un-used resources. This leads to that “Oh no!” bill from a cloud provider. It’s the one you get when you forgot you had provisioned resources and stopped using them. If your application workloads aren’t continuously right sized, your costs can rise.
What’s more, a lot of people don’t realize that cloud providers update their resources. Newer resources actually cost less. But if you simply maintain a static operational model, you could be paying more without understanding why.
Q: What can organizations do to reign in cloud costs?
A: Taking advantage of the cloud and controlling cloud costs requires a shift from a data center mindset to a cloud operations perspective. IT teams need to equip themselves with an application resource management solution that gives them the ability to optimize their on-premises applications before they move to the cloud. The solution needs to keep right-sizing application resources once they’ve moved workloads to the cloud to ensure performance and keep costs in line.
In addition, it just isn’t humanly possible to keep up with all the dynamically changing parameters in either a cloud or hybrid cloud environment. IT teams need an application resource management tool that provides the intelligence to deliver the visibility, insights, and recommendations to continuously automate resource optimization at the lowest cost. This lets you avoid creating headcount to build spreadsheets to track and adjust resource management in the cloud.
Finally, beware of the entire picture. Connect your application resource management solution to an application performance management tool, such as Cisco AppDynamics, Dynatrace, Datadog, or New Relic. This gives you visibility to know what apps are doing so that adjusting resources doesn’t impact application performance.
Q: How does Cisco help address these challenges?
A: Cisco’s approach to controlling cloud spend is built around workload optimization across on-premises, public cloud, and edge environments. Our solution, Intersight Workload Optimizer, focuses on optimizing application resources to ensure application performance by connecting all the elements in your infrastructure through what we call an integrated supply chain. This includes everything from bare metal, VMs, and containers to the dynamically changing parameters from your cloud provider and the applications themselves.
Because your application resource demands are constantly changing, IWO assesses them in real-time (both on-premises and in the cloud), then it makes recommendations on how to adjust workloads to provide computing resources when and where applications need them at the lowest cost possible.
Q: What problems does IWO solve?
A: IWO solves several problems. First, whereas cloud providers’ tools can only see cloud resources, IWO gives you complete visibility across your on-premises, cloud (AWS, Azure, and GCP), and edge infrastructure, showing the interdependencies of your computing resources across your entire stack.
Second, IWO integrates with APM tools like AppDynamics, New Relic, and Dynatrace to show you what’s happening in your applications on top of the resource layer. It recognizes when application issues are tied to workload resources and both alerts you to potential issues and provides recommendations to prevent them from happening.
Third, IWO helps you start optimizing resources and cost at the beginning of your cloud journey. Whereas public cloud providers can’t reach into your on-premises environment, IWO helps you assess how your workloads are running in your data center before you move them to the cloud. This sets you up for success before cloud migration rather than surprises once you get there.
Fourth, because IWO understands the interdependencies between your computing resources across the entire stack, it surfaces insights and alerts about where application performance may become sluggish or result in application downtime. Then it makes recommendations on how to adjust resources to prevent performance issues. It optimizes workloads to ensure continuous application performance so your team doesn’t have to operate in firefighting mode.
Fifth, IWO makes recommendations based on real-time cost factors from cloud providers, eliminating the need for your team to track constantly changing information that determines cloud costs at any point in time. It recommends the resource placements that meet dynamically changing application requirements at the lowest cost.
Finally, IWO lets you unlock the elasticity of the cloud by automating actions to continuously optimize workload resources. This ensures your applications get the resources they require based on both increasing and decreasing requirements as well as the lowest cost for cloud resources to meet those needs at any given point in time.
Q: What kind of results are IWO customers seeing?
A: IWO customers who are using one or more public cloud providers are typically reducing over provisioning by 20-30% and seeing reductions in cloud spend by 15-20%. In hard numbers, one customer lowered its cloud spend by $60,000 per month implementing actions recommended by IWO, and IWO identified $6 million in additional savings with automated optimization actions across memory, VM capacity, CPU cores, and wasted storage.
Even customers who are using public cloud as an incubator for their development environments are reducing their cloud spend. One customer reduced its AWS spend by $80,000 per month in their development environment through optimization.
Resource optimization and related cost savings also result in the data center. Customers that use IWO to assess their on-premises workloads are seeing 40-50% in cost savings when they refresh their data centers. They see that they don’t need to buy as much capacity as they thought they needed to meet workload requirements. Understanding of on-premises workload requirements is critical for anyone who needs to upgrade their data center infrastructure.
Q: What recommendations do you have?
A: The time is now. Take action. You can prevent surprise bills for over-spending in the cloud, and you can be smart about optimizing your cloud resources on an ongoing basis to truly take advantage of cloud elasticity. What’s more, by getting visibility into workload optimization across your entire infrastructure you can potentially uncover additional cost savings in your data center.
IWO is a SaaS solution and you can optimize your public cloud workloads for free for 45 days using the IWO trial. All of IWO’s features are available in the trial, so you can also explore cloud migration planning, which is the next topic in the blog series.
Reducing cloud spend is a top priority, and we’re here to help.