How Did My Cloud Bill Get So High?

Choosing the best workload placement is no longer optional, but you must consider the business value, cost and risks first

Ian Botbyl
Ed Baichtal
How Did My Cloud Bill Get So High?

Deciding where to deploy your hybrid IT workloads is complex. Do you choose the public cloud, colocation, edge, on-premises or even Bare Metal as a Service (BMaaS)? And what factors must be considered in your decision-making process? There’s cost, agility, business value, speed, reliability, compliance, customer experience, data protection, disaster recovery and more. If you choose to simply take a lift-and-shift approach to migrating workloads to the cloud—instead of creating a holistic migration strategy that fully considers all key criteria for cloud success—then cloud repatriation may be in your future.

Post-cloud deployment (also known as cloud repatriation) is the migration of workloads from public cloud to local infrastructure environments, typically either a private or hybrid cloud deployment model that can include BMaaS single-tenant compute and storage. But why move away from cloud after spending the time, energy and expense of getting workloads there in the first place?

For some businesses, a post cloud deployment is strategic and temporary as they seek to optimize their cloud and on-premises environments. Others are moving back on-premises for more complex reasons, sometimes rooted in poorly constructed migration plans with vague corporate objectives. Businesses will need to monitor workload performance continuously to ensure ongoing optimal placement.

In this article, we’ll look at the common reasons why the lift-and-shift approach to cloud migration fails, and why some organizations are intentionally moving certain workloads back to physical infrastructure.

In the early days of cloud computing, lift and shift was a common approach that enterprises used to migrate apps and associated data to a cloud platform. It meant replicating on-premises apps in the cloud to avoid costly, time-consuming redesigns. While it’s usually best to refactor an application as part of a migration, sometimes organizations need to do so retroactively—especially when time to market was their biggest priority for moving to cloud.

Assessing when migration to the cloud will add value

From the start, public cloud providers positioned lift and shift as one of the fastest and least-expensive ways for organizations to begin shifting from a CAPEX model to an OPEX model. Turns out this was an oversimplification of what it takes to migrate workloads to the cloud and the associated costs. Why? Just like any other initiative, companies need to be strategic about assessing when it’s more efficient and cost-effective to place workloads in the cloud rather than keeping them on-premises. And, once you do migrate workloads to the cloud, it’s not a “set it and forget it” scenario. Enterprises need to regularly analyze their cloud spend and ask the cloud providers to demonstrate TCO across the entire platform.

It’s rare that high cloud bills can be blamed entirely on cloud providers. It’s more likely that, prior to migration, enterprises didn’t refactor applications and prep data to run efficiently on the new cloud platforms. Also, they can’t take advantage of cloud-native capabilities that allow workloads to function efficiently, such as auto-scaling, security and storage management.

Just like the digital tools you use in your daily life, cloud is not an either-or option. Not all workloads and hardware are compatible with every vendor. Another drawback is that many niche cloud providers lock customers into long-term contracts. If you change your mind before the term expires, you can expect to pay a hefty early termination penalty fee.

Redirecting misplaced workloads

Lifting and shifting can be compared to moving an outdoor plant from one area of your yard to another. Locating a plant in a new environment with different conditions, such as more or less sun, can affect whether it will thrive or die.

Likewise, an IT project that started on-premises or on a legacy system might not perform as well in a new environment; there may be a mismatch with the handling systems, or the data sets may quickly outgrow their environment. In such cases, resource-intensive apps may need to be redesigned from scratch as cloud-native apps to eliminate performance and latency issues. Many of these issues could have been avoided, but the allure of migrating to cloud quickly proved to be too strong.

BMaaS can serve as a life preserver for migrating workloads off cloud onto a more cost-effective platform. This is because consumption of bare metal infrastructure is billed differently than a traditional cloud service provider. Let’s look at a real-world example of how an Equinix customer solved their problematic workloads with BMaaS.

Deploying to bare metal solves three customer challenges

Our customer’s internal IT team and studios needed to move fast, but their legacy IT teams lacked agility to keep up.​ Other challenges included networking issues with a migration from on-premises to cloud, needing to position storage in closer proximity to their services, and an extended migration due to technology limitations. When the customer presented these challenges to Equinix, they discovered how digital services such as Equinix Metal® single-tenant compute and storage, Equinix Fabric® software-defined interconnection and Network Edge virtual network functions could be the solution they were seeking.

The customer started with a development and test environment to assess the compatibility of running legacy workloads on Equinix Metal hardware and using Equinix Fabric interconnection to connect their physical and virtual infrastructure.​ They validated what it would take to move workloads over to Equinix Metal and tested the production solution live cutover during an on-site workshop.​ The customer then moved to the full design stages of what the footprint would look like before production.​ The team reported the following results from running workloads on Equinix Metal:

  • Workload-optimized hardware ​
  • Flexibility and time to market​
  • Proximity and performance ​
  • Ease of deployment​

Equinix Digital Services Deployed on Platform Equinix

Solving capacity and multi-region hardware issues with bare metal

Our customer required capacity for its new product in the Europe, Middle East and Africa (EMEA) region. The primary goal was to capture revenue in new and existing markets for a fast-growing product.​ Their infrastructure team was unable to provide capacity due to the product’s unique network architecture requirements. They needed to support multiple product teams with differing architectures.​ Supply chain delays added to the complexity of upgrading their infrastructure.​

The customer team attempted to use a cloud provider to host their infrastructure, but the vendor was unable to provide the necessary capacity.

Instead, they switched to using Equinix Metal, which simplified the process. The team decided to test and validate that they could run their workloads on Equinix Metal servers and use the network topology. ​The team reported the following results:

  • Proximity to their own Equinix deployments in Frankfurt and Amsterdam​
  • Access to physical servers ​
  • Time to value, due to how fast Equinix delivered servers
  • Ability to scale as capacity requirements​ expand
  • Global consistency and repeatable​ deployments
  • Integration of physical and virtual infrastructure​

Running Workloads on Equinix Metal on Platform Equinix

Move workloads to the cloud—selectively

So far, we’ve made the case for why migrating workloads and storing data in clouds may not be the most cost-effective choice for enterprises. The following use cases are some examples of when placing workloads in the public cloud may be the right choice:

  • Supporting remote workers’ use of mobile devices and applications by providing reliable access to key corporate services, documents and data from anywhere
  • Facilitating cloud-based collaboration and content management with remote access to enterprise applications such as Microsoft Office 365 and Google G Suite
  • Accommodating demand for videoconferencing services with their variable usage patterns and extensive networking bandwidth requirements
  • Hosting scale-out applications that support spikes in demand and variable usage
  • Deploying cloud-based disaster recovery solutions that support failover requirements on a pay-per-use basis
  • Setting up business continuity solutions that take advantage of the inherent redundancy of cloud environments and offer greater availability guarantees

The point here is that choosing the right workload FIRST is crucial. Don’t migrate to the cloud just because of the perception that “cloud is cheap.” Organizations that move these types of workloads to the cloud can consider the migrations a launchpad to becoming cloud native. Now it’s time to consider transitional solutions such as BMaaS or building out colocation to support your cloud migrations and repatriations.

Solving cloud complexity and lowering costs

BMaaS single-tenant compute and storage can serve as a central zone between colocation and public cloud. Building and connecting your digital infrastructure is becoming increasingly complex as it expands to meet various business objectives. Your infrastructure needs to:

  • Seamlessly connect multiple geographies, customers, partners and technologies
  • Support your choice of leading DevOps tools
  • Allow you to control cost and performance

BMaaS gives you the choice and control of physical hardware with the low overhead of cloud. Deploying compute and storage on BMaaS helps you avoid the complexity and overhead of public cloud platforms. Migrate virtual machines and workloads in and out of the cloud on-demand. Move workloads to the public cloud to reduce strain on on-premises resources when you need to create and destroy high volumes of test environments. Develop and deploy applications designed to integrate with on-premises applications and access cloud-native services.

Businesses use Equinix Metal to activate infrastructure globally, connect it to thousands of partners in the world’s largest technology ecosystem and leverage DevOps tools to deploy, maintain and scale their applications to create digital advantage.

Integrating your on-premises infrastructure with other digital services such as Equinix Fabric and Network Edge on Platform Equinix® can help you extend your network, storage and compute capabilities beyond a single location. You’ll be able to place distributed digital infrastructure where you need it, with no time-consuming and CAPEX-heavy buildout required.

To learn more about how you can build the optimized digital infrastructure you need to thrive in the digital economy, read the Leader’s Guide to Digital Infrastructure e-book.

Avatar photo
Ian Botbyl Former Senior Manager, Product Marketing
Avatar photo
Ed Baichtal Principal Product Manager
Subscribe to the Equinix Blog