Cloud Predictions for 2022

Now that we have officially wrapped up an excellent conference at AWS re:Invent and 2021 is coming to a close, I thought it would be a great opportunity to start looking towards the future and generate some hypotheses on how the cloud computing space will continue to evolve, with five key predictions. 

These predictions are largely centered around the ubiquity and configurability of the cloud, the maturity of the true multi-cloud service fabric, meaningful adoption of mission critical workloads in the cloud, the intelligent cloud, and the clean cloud.

Let’s dive in!

Prediction #1: The rise of the “configurable cloud” across physical, regional, and industry dimensions

In the early days of public cloud, regions and their underlying physical data centers were highly centralized in hyperscale facilities. While that still remains true, the Big 3 Cloud Service Providers (CSPs) have started to react to specific customer needs rather than the original incarnation of “commodity cloud” and become more of a ubiquitous ecosystem that can be configured to any customer needs. Some great examples include:

• Physical requirements: New AWS Local Zones deliver lower latency in Chicago and New York City where low and ultra low latency in a must, especially in the Financial Institutions sector
• Regional requirements: the regional coverage globally across all CSPs has significantly increased as well as regional-specific offerings, such as the Microsoft 365 Multi-Geo offering from Microsoft Azure
• Industry requirements: the rise of industry clouds that acknowledge and cater to unique industry and sector requirements (e.g. regulatory); a great example here includes the recently announced Microsoft Azure industry cloud offering 

A good example of industries with very bespoke physical, regional, and industry specific requirements includes the Financial Institutions and Public Sector, primarily due to mission critical performance requirements and highly sensitive workloads & data. Some of the specific deployment models to solve for industry requirements can include the following:

Figure 1: Latency considerations for cloud deployments

… you can (1) connect your existing data centers to the CSP environments you are adopting, selecting the optimal physical proximity and regional configuration; (2) integrate with existing cloud fabric with existing or new co-location facilities; or (3) possibly even considering bringing the cloud to you via cloud region and data center co-location that physically reduces or even removes physical latency that is typically introduced in a hybrid or multi-cloud deployment. 

This flexibility is allowing for ease of multi-cloud deployments, and starting to abstract away the challenges of adopting multi-cloud. Additionally, it is providing a continuum of choice to customers that are debating between going cloud-native, cloud-agnostic, or somewhere in between:

Figure 2: an increasing focus towards cloud agnosticism by many organizations

Overall, the key prediction here is that the “matrix cloud” will be born, where once uniform commodities will now be highly configurable to allow customers to slice the cloud any way they want horizontally (physical / regional) or vertically (industry), which will also in turn allow for a more agnostic and flexible approach to public cloud adoption. 

Prediction #2: The establishment of true, end-to-end multi-cloud service fabric

More and more, customers who adopted single cloud models, or even multi-cloud by accident, are starting to refresh their strategies towards a more deliberate multi-cloud strategy. Additionally, all of the Big 3 Cloud Service Providers have officially shifted their strategies towards multi-cloud. That being said, it is still early days as far as service maturity and when you look at the cloud data & service fabric components below:

Figure 3: Cloud data & service fabric 

… some of the areas of relative maturity include compute, where multi-cloud enabled container offerings and CSP-native as well as CSP-agnostic PaaS offerings have been generally available for some time now. Some examples include GCP Anthos, RedHat OpenShift, AWS Outposts and VMware Tanzu. The biggest area of relative immaturity is mainly in the orchestration and networking space, where most CSP-native services are vastly different and require a high degree of integration if not a completely agnostic layer that resides on top of each CSP environment for a more homogenous ecosystem (capabilities from GCP Anthos also include the orchestration space). My hypothesis is that this space will continue to evolve and mature to where cross-CSP orchestration is significantly more abstracted and automated than it is today.

Another area that will continue to mature lies within cloud networking advances. In building upon dedicated network ecosystems for turnkey cloud connectivity (such as AWS Direct Connect), we recently witnessed an AWS Cloud WAN announcement which bring a new managed WAN service to seamlessly operate & maintain a multi-cloud environment that includes on-premises. 

Anthos competitive services & equivalency will take rise. GCP Anthos was the first to market in a public CSP providing a service out of the box that natively extends to both on-premises as well as other public cloud environments. The reticence towards true multi-cloud enablement is starting to shift, and i believe that based on recent developments all CSPs are embracing this operating model and expanding their service stack to reduce the friction to enabling multi-cloud. The battle will no longer be about being the only cloud, but rather the “anchor cloud” in a customer’s multi-cloud operating model. 

Prediction #3: An increase in velocity of mission critical workloads towards public cloud

While recent thought leadership suggests that highly regulated industries such as Financial Institutions which take a unique and much more deliberate path to the public cloud, the pivot has already started with customers looking to modernize core systems and consider public cloud for mission critical workloads. 

The defense and national security space is a key example. For example, the Department of Defense has shifted its cloud strategy from an enterprise to a warfighting capability centric strategy and cloud. Additionally, they have shifted the way they acquire and procure these capabilities to match its strategy. The “mission critical cloud” has started in a big way. 

While Financial Institutions are another mission critical industry that is extremely highly regulated and in our view are on a very unique and deliberate path to the cloud, some recent cloud announcements make note that this is evolving. A great example includes NASDAQ’s announcement at re:Invent to shift its options market to AWS, forming a partnership to build the foundation of the next generation of capital markets.  Additionally, Goldman Sachs also announced an AWS collaboration to create new data management and analytics solutions for Financial Institutions.

A great reduction of the barriers and friction here includes the AWS re:Invent announcement of its new Mainframe Modernization offering, as the majority of these legacy systems reside today on mainframe systems and database environments. 

Prediction #4: Evolution of a rich ecosystem for AIOps and MLOps as a Service (AIOps-aaS)

We have learned previously that the cloud could be smarter and more intelligent than it is today as legacy ways of working and operational constructs can and have already replicated themselves in the cloud for many organizations. A fast-growing set of tools and approaches to combat these problems lie within AIOps (also referred to as MLOps depending on vendor and use case), which is a method of utilizing big data, advanced analytics and machine learning to enhance and automate IT operations & monitoring, enabling intelligent operations of your IT environment. While it is early days, IT leaders are starting to employ machine learning (ML) technologies within their own functions to effect operational efficiencies and cost reduction for enterprise technology. In the past few years, a number of vendors have begun designing and developing powerful analytics tools to address the particular challenges that IT personnel face in managing, updating, and running IT hardware and software across the enterprise

CSP-native offerings include partnerships with independent software vendors (e.g. AWS and Moogsoft) as well as CSP services that by nature can be designed and configured to achieve AIOps outcomes (e.g. GCP MLOps). CSP vendors are increasingly developing services in-house and onboarding partner technology vendors to create an ecosystem that is highly supportive of enabling AIOps and MLOps use cases.  

Product focused companies like DataRobot are also expanding their platforms and offerings that are cloud-native, with the recent release of DataRobot AI Cloud as a great example. 

The prediction here is that a combination of CSP-native and independent software vendor offerings will make it extremely easy for customers in the cloud to latch onto and bolt on 

Prediction #5: The Clean Cloud serving as an accelerator to sustainability

The clean cloud is critical to sustainability and national security imperatives. Amidst growing confidence that government can fuel a green recovery, the Biden administration set a bold vision for a climate change and infrastructure agenda. What may not be apparent at surface level is that emissions from information and communications technology are significantly contributing to the carbon footprint and broader climate change. If left unchecked, the IT related contribution will surpass the entire aviation and shipping industries and total 5.5% of the world’s carbon emissions by 2025. Over the past two decades, Google’s data center program, which serves its enterprise and cloud offering, has set the standard with its approach to data center efficiency at industry leading benchmarks.  Moreover, it measures progress against a goal of having operating data centers and campuses on carbon-free energy, 24/7, by 2030. CSP vendors have quickly become the market leaders in Green IT by building enterprise data centers that meet industry leading sustainability goals and exposing the same infrastructure as services to its commercial cloud customers. 

The prediction here is that as many commercial entities and cloud customers make commitments to sustainability goals, those same goals will be expected upon their in-house infrastructure and the cloud environments they consume. The CSPs that provide the best path to sustainability will likely have an advantage in helping cloud customers reduce assets that are not supportive of sustainability goals in favor of consuming cloud services that are asset light and by nature are green IT resources. 

Conclusion.

Let’s check back in a year and see which of these show follow through! One thing for sure is that the cloud is becoming increasingly entrenched in the technology strategies of the world’s leading commercial and public sector organizations, and will only become a bigger part of the foundation of the next generation of enterprise and mission critical technology. 

Photo by Caspar Camille Rubin on Unsplash