Cloud at the crossroads

A blog post by David Linthicum, Managing Director, Chief Cloud Strategy Officer

For the past decade, the topic of digital transformation has dominated the conversation among CIOs, and it’s no surprise that cloud is at the center of these conversations, as cloud is at the core of most business and digital transformation efforts. Also, as artificial intelligence (AI) is becoming more accessible than ever before, it’s important to remember that cloud makes AI and other advanced capabilities more accessible and affordable than ever. But we still have a long way to go with ensuring that companies get the required returns on their cloud investments. Solutions such as FinOps and cross-cloud abstraction and services will play a critical role in boosting cloud success, and cloud’s long-term future is bright.

We’ve come a long way with AI and other tech advances

Artificial intelligence was developed in the 1950s but lay mostly dormant for decades, primarily because it was cost prohibitive to implement. In the past decade or so, however, cloud computing has made AI accessible and cost effective, so companies can harness its power and punch above their weight to compete. These advancements have made true digital transformation a reality.

What may be most indicative of how far we’ve come, though, is the ease with which these AI and machine learning (ML) projects may now be implemented. Today, it’s possible for anyone to use Python and libraries on the internet to create compelling and complex ML applications. The fact that developers can build something so powerful in mere weeks in some situations makes transformation not only easier but also more exciting. What’s more, in addition to faster implementation, these complex projects have also become more affordable. For example, the development of web APIs has made integration possible at a fraction of the previous cost.

But there’s room for improvement—especially with multi-cloud

As it’s becoming easier and more cost effective to scale digital transformation efforts, multi-cloud architectures have become almost de facto for large organizations. These companies typically have one or more commercial clouds in use, and by leveraging SaaS applications from vendors (who, in turn, use other commercial clouds), there’s more distributed data than ever before.

This disparate data environment makes it challenging to consolidate data for analysis and insight—which is essential for AI/ML efforts. As companies attempt to reconnect this distributed data, they may face a “multi-cloud reckoning” that will force IT leaders to seek better ways to manage their multi-cloud environments, optimize their efforts and costs, and access the data they need for their AI/ML projects.

Cross-cloud services and FinOps: Key factors in optimizing cloud

To help rein in multi-cloud complexity and increase cloud value, two emerging technologies will play a key role: cross-cloud services and FinOps.

A cross-cloud approach combines all cloud assets within an organization—from APIs to SaaS products—and makes it easier to integrate, orchestrate, and combine cloud data and resources to securely leverage and create new solutions. Put simply, cross-cloud is an abstraction layer that sits above multi-cloud deployments to offer more effective management capabilities under the “single pane of glass” concept.

We also anticipate that the use of FinOps will grow enormously because of its ability to focus observability efforts on managing and controlling costs. With FinOps, finance and DevOps teams use observability data to collaborate to better understand where and how their cloud spend is generated.

FinOps also uses AI and ML to automate routine tasks and gain deeper insights into cloud costs. By embracing FinOps, enterprises can make data-driven decisions that can help identify and reduce wasted cloud spend and optimize cloud costs—which can enable them to increase the value of their cloud investments.

Demonstrating value from cloud investments is paramount

For the past 15 years, the public clouds have gone through multiple maturity stages—from normalization of the market to commoditization of the cloud process and cloud services. In the current market, there are now enough feature-rich services available from cloud vendors to accommodate most applications that companies want to migrate—or build cloud native.

We’re reaching a point of feature saturation, however, so cloud providers are looking to optimize existing features like storage, databases, and processes; and they’re leveraging containerized applications that add more value for the business. The high cost of cloud services is also forcing leaders to scrutinize their cloud costs more. This increased pressure is impelling cloud vendors to find ways to demonstrate value for customers.

The bottom line is that no matter what platform you use—private cloud, public cloud, legacy systems, or hybrid architectures—the main objective is to create an architecture that can provide the greatest return for the business. And, as CIOs are concerned about ballooning cloud costs, optimizing those costs is high among their priorities. This has companies looking for ways to reduce their cloud bill without reducing service level while getting the most out of their cloud investments.

What the long-term future of cloud looks like

Tracking the R&D and innovation rates of commercial clouds is difficult due to hyperscalers’ rapid innovation rates, but there are some standout trends we’re excited to see in 2023, such as the emergence of streaming data to fuel the use of real-time applications that can give customers answers based on real-time events. Database choice is another feature that hyperscalers are working overtime to provide.

Another exciting differentiator is custom silicon. In the not-too-distant future, specialized workloads will have chips designed to optimally run them, which won’t necessarily reduce the cost of cloud services per se but will provide a lot more efficiency that could reduce costs in the end. 

Finally, developer education is going to be very important in the near term. With the increasing number of services, understanding how best to write them together takes time, and that’s slowing adoption. The introduction of architectural patterns and reference architectures will, hopefully, empower developers and allow them to absorb these new services much faster.