The rise of multicloud architectures in enterprises is a reality, with three-quarters of businesses in CCS Insight’s Senior Leadership IT Investment Survey, 2022 using more than one cloud provider, compared with only 55% in 2021. The number of cloud platforms in use is also growing: on average, businesses have 2.5 cloud providers, up from 1.9 in 2021.This increased shift to multicloud strategies, alongside the growth features in each platform, can make managing the complexity of an enterprise IT estate a daunting proposition.
Cloud maximalists insist that single-cloud architectures are more reliable, but thoughtful deployment of workloads in a well-architected multicloud deployment can provide an optimal experience — enterprises can choose the features, infrastructure and venue that best fit their workloads. To accomplish this, a common control plane can assist in architecting a hybrid or multicloud deployment and reduce the toll on developers, operators and security practitioners in day-to-day maintenance.
Reducing the Complexity of Multicloud Architectures for an Optimal Experience
Multicloud deployments can deliver advantages to enterprises in a variety of circumstances, depending on:
- The existing workloads involved
- The budget allocated for cloud services
- A leading cloud preference, rather than just a single supplier
- The timing of the refresh cycle for on-premises infrastructure
- Progress in migrating workloads on the infrastructure to public cloud platforms
For example, organizations with existing on-premises Microsoft Exchange or Windows Server installations may find Microsoft Azure a more comfortable fit as Microsoft’s licensing terms make Windows-bound workloads more affordable on Azure than on competing platforms. In other circumstances, organizations with artificial intelligence (AI) or machine learning workloads may want to use high-performance silicon such as Amazon Web Services (AWS) Trainium or Google Tensor, which are exclusive to those cloud services.
More generally, enterprises with data sovereignty requirements or concerns about being locked into one supplier may prefer a multicloud deployment to increase resiliency and agility. Particularly for businesses, multicloud environments can organically occur as the result of mergers and acquisitions, often requiring a labour-intensive audit of the inherited assets.
Although this sounds agreeable in theory, multicloud deployments are complex systems that require careful design and a comprehensive toolset to reduce wasteful resource allocation and the toil that IT professionals experience in building, maintaining and securing these assets.
Achieving Cloud Coherency Using Multicloud Services
In a way, all cloud platforms provide the same resources: computing, networking and storage. The APIs, infrastructure and interfaces are unique in each platform, making the task of managing multiple clouds more difficult as more are added. Abstracting these to a single common interface can aid in reducing complexity, but abstraction alone simply reduces each cloud to the lowest common denominator, obscuring the unique attributes of each cloud platform used to assemble an optimal solution.
With a multicloud service, workloads and the infrastructure they run on can be managed — in a cloud-agnostic way — throughout the entire application life cycle. This allows developers to use their preferred framework without needing to manually accommodate every potential execution environment. This, in turn, reduces the strain on operations professionals, as workloads can be managed and monitored in the same way over public cloud platforms and on-premises infrastructure.
Others can also benefit from a well-architected multicloud environment. Security professionals can more easily apply consistent security policies over multiple cloud platforms by relying on multicloud control planes, and financial operations professionals can correlate resource allocation and spending on cloud architecture with the responsible business units. With tools that accommodate the needs of the different roles in an enterprise, cloud complexity can be transformed into cloud coherency.
Planning Your Multicloud Software Ecosystem
Achieving cloud coherency requires a software stack that is feature-rich and actively developed, ensuring that newer functionality in cloud platforms can be employed in a timely manner. For example, VMware Aria provides up-to-date support for bridging cloud platforms and on-premises infrastructure, with assistance for computing, network, security functionality and storage, as well as cost management and performance optimization features through CloudHealth.
In addition, VMware Aria integrates workload observability features, including a “near real-time” map of cloud assets and the applications running inside those assets. This is particularly beneficial to organizations auditing IT assets inherited through acquisitions, though equally useful to traverse shadow IT activities. VMware Aria Business Insights, similarly, is an AI- and machine learning-guided event-correlation service that highlights significant events that could result in service degradation or outages, informing troubleshooters of problems as they develop.
In comparison, open-source software provides functionality for computing, network management and storage, but aspects such as cost management, performance optimization and security policy compliance are generally commercial value-added options.
OpenStack services provide many of the elements needed in principle, although work on API compatibility has slowed in recent years. Support for AWS Elastic Compute Cloud remains relatively current, but connectors for Google Compute Engine and the multiservice OpenStack Omni have not received commits in several years. Other functionality can be supplied with an assemblage of open-source software, such as a stack of Elasticsearch, Logstash and Kibana, commonly used for observability.
Terraform, an infrastructure-as-code tool, has also made inroads among enterprises for virtual infrastructure management. HashiCorp sells Terraform Enterprise as a value-added product over the top of the open-source version, but third-party terraform automation and collaboration software including Scalr, env0 and Spacelift are also available.
Getting Ahead in the Cloud, Without Getting Ahead of Yourself
The time dedicated to platform and infrastructure engineering versus the time spent on the products or services sold a company is ultimately a balance of priorities for the individual roles in an organization. There is complexity inherent in relying on an assemblage of open-source, internally developed and commercial software to manage multiple clouds, and additional costs of paid support and infrastructure power an orchestration layer for your multicloud estate.
In environments without existing cloud systems, this endeavour would be easier to undertake — thoughtfully building a new model is easier than gluing together existing disparate systems into a unified operating model. That said, these environments are few and far between, making integrated solutions such as VMware Aria an attractive choice for organizations seeking a turnkey solution for managing IT assets for multiple clouds.