One of the main drivers for moving workloads to the cloud is the ability to reduce costs. However, once migrated, cost reduction is also one of the biggest pain points.
That's because the nature of cloud computing is its scalability, so users naturally want to use more and more compute power, observed Owen Rogers, research director of the digital economics unit at 451 Research. While the cloud might start off cheaper, those costs ramp up as consumption increases.
This dynamic presents a good opportunity for partners to help customers with cloud cost optimization and management, on an ongoing basis, especially when it comes to hybrid clouds.
Hybrid cloud is the best of both worlds, Rogers said. For workloads that need highly predictable performance and are likely to remain fairly constant in their capacity needs, however, private clouds can present the better option. 451 Research's Cloud Price Index found that private clouds can be cheaper than public cloud when utilization is above about 60% and labor efficiency is above about 600 virtual machines (VMs) per engineer.
But with a hybrid cloud computing model, "buyers can get best value for [their] money for constant workloads, with the freedom to burst in the public cloud when needed," Rogers said. "Partners offering a hybrid solution have a good differentiator here over the hyperscalers."
For all customers, regardless of size, hybrid cloud cost optimization is a concern, said Craig McQueen, director of the Microsoft practice at managed services and consulting firm Softchoice. Customers have varying levels of awareness that by implementing some changes they could actually be saving money, he noted. "Larger customers, additionally, have a concern about being able to allocate costs to business units."
Although there is reporting software for hybrid cloud cost optimization recommendations, "what we see is customers are slow to take action on the recommendations," McQueen said. "They are either unsure how to implement the changes or silos within their organization make it difficult to find the person who should make the changes."
Partners can help clients prioritize and implement the hybrid cloud cost optimization recommendations, such as right-sizing VMs or automating use of a resource through scheduling, he said. "Customers also struggle with implementing a cost allocation strategy. Partners who have an approach and tools to support that need can bring good value."
Pain points of hybrid cloud cost containment
Rogers sees two broad issues when it comes to controlling hybrid cloud costs. The first is making sure that money is not wasted -- for example, by leaving VMs running when they are no longer doing anything. The second is ensuring that a customer is using the cheapest combination of resources for their use case.
Hybrid cloud cost containment is more of an issue for small and medium businesses (SMBs), said Jim Lippie, general manager of cloud computing at Kaseya, an IT management software vendor that works with managed services providers. Kaseya acquired Unigma, a cloud management software vendor, in May 2017.
Enterprise organizations, on the other hand, have much more mature cloud deployments, "so it's not new to them," Lippie said. They tend to have a better handle on the full scope of what they're spending and where they're spending it. "Most SMBs are looking at cloud from a new perspective and a new lens" and are not as familiar with the associated costs and how to reduce costs once they have moved to a hybrid cloud computing model.
Cloud computing often starts within the development team of an organization that has a need to quickly get software deployed, and the team bypasses the infrastructure group -- which has greater awareness of ensuring IT policies are followed, McQueen said.
Problems can arise if the lifecycle management of those resources is not in place, he said. Like Rogers pointed out, McQueen said developers may start deploying cloud resources for development and testing but then keep them running even when they are no longer needed.
"They continue running and costing the company. Additionally, companies often implement a tagging strategy for resources; that is, ensuring that cloud resources are labeled with things such as a cost center when deployed, McQueen said. "If this isn't done, it makes it difficult to implement billing allocation."
Tools for hybrid cloud cost optimization and management
There are new tools in the market, such as Spotinst, DivvyCloud and ParkMyCloud, that can help ensure a customer is using the most cost-efficient resources for their use case, Rogers said.
Owen Rogersresearch director of the digital economics unit, 451 Research
Kaseya's Unigma tool includes a module that lets organizations look at specific configurations within the three large cloud providers: Amazon Web Services, Microsoft Azure and Google. The providers themselves are "never going to come to you and say, 'You have an oversized instance with us and we can reduce your CPUs,'" Lippie said. Not surprisingly, he said, the Unigma module is very popular with enterprises.
Softchoice bundles a tool as part of its managed services offering but believes "a tool alone won't solve the problem" of cost containment, McQueen said. "The customer needs a good strategy for categorizing resources, setting up billing centers and actioning cost optimization recommendations. We find only by having the cost best practices blended with a tool, do customers get the results they need."
Helping with hybrid cloud cost containment
Ongoing hybrid cloud cost optimization is a key opportunity for partners, Rogers said. "Buyers don't want to mess around with constantly optimizing workloads. A managed offering that takes care of this lets the buyer consume as needed, without worrying about wasted cost-saving opportunities or spiraling costs." He stressed that this process needs to be a constant and ongoing value-add.
McQueen also concurred that taking the ongoing action on hybrid cloud cost optimization recommendations is a core best practice for partners.
Other best practices he suggested are:
- Create a tagging model and strategy for ensuring it is used.
- Identify cost-saving opportunities.
- Implement a reporting system for billing.
- Automate resources so they are only used when required.
Lippie agreed that providing a constant, ongoing value is a must. The value a partner can have is to monitor the client's cloud environment to ensure it is the right-sized infrastructure and the right resources are used.
He said partners should report to the client on what they're utilizing on a monthly basis. Partners should also tell the client that if there is a solution where the client can cut costs, "'We'll come to you and say, 'Hey, we can bring this down X dollars per month because you're only utilizing this amount of CPUs and storage.'"