The Old Managed Service and the Sea: Cloud Economics Trends in 2020

Introduction

Cloud in 2019 can be summed up in one word: volume. More services, offers, tiers and payment options have caused the number of cloud price items for sale from the dominant hyperscalers spiral to over two million. The sea of options is only likely to get deeper and choppier. As we discuss in the Cloud Price Index's annual report available to CPI and Cloud Transformation subscribers, 2020 will see enterprises look to optimize and control complexity with intelligent tools and managed services

The 451 Take

Once able to potter around safe harbors of virtual machines charged by the hour, enterprises must now navigate a swirling sea of great complexity. Evaporation here leads to new rainfall, which restocks the sea with new products, services and options. Sometimes the swell may rise and fall as prices change temporarily, and occasional turbulence will sweep price changes across a small patch of cloud services. But the sea is getting deeper as service providers develop more capabilities, and also choppier as new tiers, generations, technologies, chipsets and pricing models are created and renewed. Although each droplet matters, the interactions between them are what dictate sailors' success or failure. Navigators must look at the bigger picture to manage this complex seascape, as this will ultimately shape the direction of travel. Hybrid cloud and multicloud provide new currents, pulling IT decision-makers in different directions; some workloads are better suited to follow one course than another. What is needed is a lighthouse – an independent, trusted entity to help plot the best route as changing conditions make navigation more difficult than ever. The old managed service must evolve to support the new complexity of this multicloud, multi-service world.

Turbulence in the Bay of Cloud

While competition among the largest public cloud providers remains intense, the market has largely moved past the price wars and one-upmanship that grabbed headlines a few years ago. Yes, the cost of basic compute and storage services is still trending downward, but not as dramatically as before. In pricing terms, the cloud infrastructure market has found its level – for the moment – with hyperscale providers able to profit while continuing to build out their portfolios to meet customer demands and offer higher-level platform services that make infrastructure choices invisible to the user. Rather than attention-getting across-the-board cuts, cost reductions are likely to be tied to a particular technology change – say the introduction of a new processor with better price/performance, or a new class of storage. This is creating turbulence on the cloud sea as enterprises struggle not only with changing prices, but also with new services that might have better value or be more appropriate for their needs – will serverless platforms eventually do away with virtual machines, for example?

As their familiarity with public cloud grows, enterprises are getting wise to the fact that they must optimize their cloud spending to minimize waste without compromising performance. Tools from cloud providers themselves as well as independent software vendors can monitor usage and set up guardrails or alerts to keep purchases within limits. In fact, several cost management and optimization startups were absorbed by established players in 2019, including RightScale and RISC Networks (purchased by Flexera), Cloudability (Apptio), ParkMyCloud (Turbonomic), TSO Logic (AWS) and Movere (Microsoft). But these players were just scraping the surface of optimizing environments based on known trade-offs, not weighing deeper choices such as swapping out one technology for a better one (e.g., replacing virtual machines with serverless).

Cloud cost and spending management used to be about performing billing forensics and tariff management against (historical) usage data. The tools in this sector have subsequently evolved (many have been acquired) to support remediation against spending levels, automated actions (stop/start, etc.) and ongoing optimization (including instance sizing and selection). The next set of enhancements will employ machine learning, analytics and greater automation to deliver 'what if?' modeling to enterprises, as well as managed services to use this intelligent tooling alongside business objectives to determine an optimum path.

But with the breadth of cloud goods and services available in the market, the key to success will be finding the right combinations of them and operationalizing them to deliver the benefits as being advertised by their suppliers – speed, scale, agility and cost.

As this re-platforming to cloud accelerates, the question is: which applications, in what sequencing are going into which venues over what time period. As the worlds of outsourcing, hosting, managed services and cloud converge, the options are growing exponentially. Best execution venue (BEV) strategies will enable users to determine which hybrid services are right for their needs. BEV strategies center on the notion that every class of IT-related business need has an environment where it will best balance performance and cost, and the IT organization should be able to select that environment as part of the general practice of IT.

451 Research's Public Cloud Price Index tracks changes in cloud service pricing and availability around the world, as well as which managed services are available and which convey the greatest value for end users. Along with information from IT decision-makers and budget holders (drawn from 451 Research's Voice of the Enterprise and Voice of the Service Provider surveys), this gives us a picture of where the cloud market has been, how it's developing and where the biggest opportunities lie. In addition to benchmarking prices on a selection of cloud services, the Cloud Price Index tracks the stock-keeping units (SKUs) available from the top five hyperscale providers: AWS, Azure, Google Cloud Platform, IBM and Alibaba.

What stood out most in 2019 was volume – more services, more offers, more tiers, more payment options – reflecting increasing depth and a choppier sea. When independent software vendors started building dashboards based on cloud APIs a decade ago, they couldn't have imagined the sheer number of public cloud services that are now available from the top hyperscale providers. At the end of 2019, we were collecting information on over two million SKUs, including more than one million for AWS EC2 instances. Beyond the storage-, compute- and database-as-a-service rudiments that dominated early cloud offerings, enterprises are now using hybrid cloud, cloud-native frameworks and innovative services like AI, serverless and the Internet of Things higher up the stack to build their applications. Rather than keeping things simple, enterprises are choosing complexity because it delivers value in the form of differentiated offerings, more efficient applications, happier customers and lower costs. Enterprises are craving more services with more granularity because it gives them the ability to add greater value to their own businesses – but this creates complexity, and a greater range of products and options to navigate through.

Figure 1: Cloud Price Index Selected Price/Service Changes 2019

Source: 451 Research Cloud Price Index

2019 in Trends

Prices on infrastructure compute and storage continue to come down as provider cost bases reduce through standardized, cheaper and more powerful hardware and greater economies of scale. As the primary battleground of the so-called 'price war,' we can expect this to continue as providers look to squeeze every cost. Over the past two years, compute has dropped a hefty 17%.

Product-line pricing allows service providers to create less performant tiers at lower costs while charging a premium for a better experience. This is already taking place with storage and hardware, and it is a tactic that we expect will continue as service providers look to give enterprises options. The Cloud Price Index has seen storage services undergo nearly 4,500 price cuts in the past year. The range of VM sizes available has increased 317% since 2015.

In newer services, price cuts are less aggressive, partially as a result of fewer options on which to cut, but also because they offer less opportunity for internal cost reduction. Because the services have yet to reach full maturity, they have not attracted enough customers to really squeeze costs on the back end. However, some price cuts are still taking place in a bid to woo customers to experiment in the hope that this will generate volume. AI and machine learning services have undergone 61 price cuts, which we think is partially driven by the need to encourage experimentation.

International expansion is also increasing complexity. The hyperscalers we track added 15 new regions in 2019, although most of them were in the US and China. Differing prices and variations in service availability from region to region create a patchwork of capabilities and costs that add to the challenge for multinational businesses.

These factors, plus some poor API discipline, is creating vast complexity in the cloud market, where products, pricing models, product variations and actual prices can change from week to week. Rather than a fixed collection of services, cloud is now a dynamic sea of change. In 2019, the CPI tracked a breathtaking 1.2 million changes in SKUs from five hyperscaler providers.

Bringing private, hybrid and multicloud into the mix further increases complexity. Enterprises want a vast selection of services because it allows them to build capabilities that differentiate their businesses, but few buyers are using the options available to them to maximum effect due to sheer breadth of capabilities and their rate of change. Not only does a wider menu of options from multiple providers add value, but it can reduce costs – by up to 32% in our analysis.

Automation and managed services are a must for enterprises, not to calm the sea but to navigate through it, finding the best combination of services at the best price to meet changing requirements, today and in the future.