Why mainframe modernization is a must with a multi-cloud strategy

22/09/2021 minute read Rob Anderson

This blog discusses the painful implications of retaining the mainframe when adopting a multi-cloud strategy, and how legacy modernization can bring unmatched agility and power to your cloud deployments.

Small companies and startups are now born into the cloud. They get their email taken care of with Office 365 or Google's G-Suite, their web presence is entirely virtual, and they typically expand infrastructure in the cloud without ever purchasing a single physical server.

The barrier to compute power is so low in a cloud vs. premises cost benefit analysis, it seems like a no-brainer. However, most companies that demand incredibly high-performance compute and massive scalability are not startups or small; they’re large and they’ve been around a long time. Many of these organizations still like the mainframe, and those with extreme aversion to change when it comes to critical systems have relied on technology bridges to carry out modernization tasks. Though they’re acknowledging the challenges that legacy code brings upon their organization, “just move everything to the cloud” sounds like an unrealistic mantra for these firms.

For example, historically, a common path away from the physical mainframe (and the expense it brings) has been to rehost mainframe applications in the cloud. This decision makes sense as part of strategic digital transformations that focus on DevOps enablement, key customer journeys, disaster recovery, or storage first. But when firms adopt multi-cloud strategies, and as the power and abstraction of innovations built for these environments (such as Secure Access Service Edge [SASE] and Kubernetes) increases, the limitations that legacy code present will grow into liabilities.


The best of both worlds

A hybrid cloud environment consists of a mix of on-premises and cloud services that are connected to allow orchestration between each in such a way that they behave like a single environment. For example, in a hybrid cloud model, permissions and identity management extend across servers that reside locally and services in the cloud without the need for replication. As such, hybrid cloud is emerging as a popular solution for larger businesses whose digital transformation strategies include cloud adoption. This is because it allows for a phased approach that's easier and less risky for the organization to digest; as opposed to a “big bang” modernization approach that moves everything from the mainframe to the cloud at once.

As firms embrace hybrid cloud adoption, they naturally progress towards multi-cloud architectures to account for strengths (or weaknesses) of different cloud providers or when they must deploy compute across different parts of the globe. This strategy builds on the hybrid cloud model with the deployment of assets into multiple cloud providers (or multiple regions of a single provider's offering), with connectivity between each to meet the look and feel of a single environment.

Multi-cloud has several distinct advantages that organizations can capitalize on, including (but not limited to):

  • Vendor-specific products
  • Vendor agnosticism
  • Simplified achievement of regulatory compliance requirements
  • Added resilience
  • Extended geographical presence


The case for SASE

While multi-cloud strategies evolve into mainstream best practice, the concept brings with it added complexity in connectivity, with VPN (virtual private network) and MPLS (multi-protocol label switching) options differing widely between carriers and cloud providers. SASE simplifies multi-cloud connectivity while increasing security by acting as a cloud-brokered, zero trust VPN-style overlay, using the internet as transport.

Since SASE is software-defined, developers can also embed network establishment into application code, which makes it easy to scale and automate cloud agnostic connectivity with the same tunable power that Kubernetes delivers to cloud deployment and container orchestration.


Embracing multi-cloud through modernization

While it's impossible to tell the future, it's clear that multi-cloud, with all the flexibility, automation, and abstraction its related products and services bring, will be a popular strategy for the world’s largest brands. As new issues (such as cloud connectivity) are addressed with cutting edge solutions that aren’t compatible with legacy systems, refactoring procedural codebases to object-oriented and cloud-native languages (such as Java and C#) will emerge as the ideal method for embracing innovation while retaining the business logic, functionality, and history of legacy systems without the shackles of outdated technology.


Further resources