Adapt or Die: Why the Cloud Computing Revolution is Far From Over

When the World Wide Web was fresh, shiny and new, we talked about it as a place that we went to, a destination we visited through our browsers. We “got on the information superhighway” and we “surfed the Web” — terms that we laugh at today.

This first era of the Web helped pave the way for incredible innovations and fundamentally changed the way we communicate. Over time, the Web has become much more ubiquitous: it’s no longer just a place we surf or simply visit, it’s part of nearly every modern means of communication and embedded in nearly all electronic devices — and the distinctions between applications, services, and ‘sites’ are nearly invisible today.

It’s also evolved from a consumer-based utility to a critical underpinning of enterprise IT. But this adoption took time, and changed slowly as Web technologies became more robust, reliable, secure and flexible. As a result, businesses now view the Web as more of an IT capability than a destination.

We are seeing a similar evolution with cloud computing. The early cloud pioneers, such as Amazon Web Services and Salesforce, created a powerful new way to deliver IT — as a utility or a service that anyone could access and consume, giving birth to new services that have become the fabric of our day-to-day lives. New technologies, techniques, and organizing principles sprouted up to facilitate and describe how to use this new model for building applications and delivering IT.

Traditional software vendors raced to figure out how to move to the as-a-service paradigm, some failing to cross the divide. Others exploded during the transition, giving way to startups pitching new and better ways to use the cloud to deliver enterprise-scale IT. This is the accelerated world we live in today, and it’s leading enterprises to rethink how and when to use the cloud for their most important workloads and systems.

High performance computing (HPC) has been equally powerful in the evolution of the cloud. For many years, HPC was relegated to the scientific and academic communities as a specialized practice for addressing some of the most challenging real-world problems. Simulations of weather, disease, financial markets, and even animal migrations have been possible using the power of HPC. Today we can decode genomes, predict typhoons, model drug interactions, and engineer advanced materials and structures by utilizing these techniques and tools.

Labs and businesses deployed HPC systems as large farms of servers, harnessed by distributed applications and scheduling technologies that were able to leverage the power of a logically unified computing system. In this sense, HPC has paved the way for sharing compute resources, distribution of work, and the interactions between applications and systems across large clusters of independent physical resources — all crucial elements of cloud-driven enterprise IT.

The patterns for software that could asynchronously interact, yet work in a coordinated fashion across a large computing cluster, play an important role in influencing many of the methods in cloud computing. Today, HPC systems are becoming even more relevant as enterprises combine Web-scale infrastructures with IT virtualization to create the underlying fabric that makes cloud computing and big data processing and analysis possible.

Resistance, velocity, change, and control

Getting to this point hasn’t been easy as many enterprises have faced a number of barriers when it comes to cloud adoption:

• Fear of Change: Like any technology trend, the “newness” of cloud computing was an early friction point.
• Skill Shortages: While organizations are ramping up, finding experienced architects to design modern cloud infrastructures, or developers capable of building cloud native applications and using CI/CD development processes at scale can be a real challenge.
• Enterprise Readiness: This is not only about the “-ities” (reliability, security, availability, etc.) but when, where, and how enterprises can introduce cloud computing to existing IT environments and delivery models.

Let’s go into that last point further. Building an all-new IT system on modern cloud architectures is not yet fully viable for many enterprises for a variety of reasons. Existing investments in IT — including depreciation schedules, licensing contracts, compliance related infrastructure and policies, and expertise — can be the biggest areas of friction for cloud adoption. CIOs must face questions about the costs, risks and benefits associated with migrating core systems. Many will never migrate but will instead be used with new cloud-based services, while others will migrate over longer periods or as technology refresh cycles dictate.

The tempo for cloud innovation is also extraordinary. It seems each week there is a new approach or technology trying to advance the field at various levels of the cloud software stack. With that come changes in configuration management, new deployment tools, new schedulers, new data services, new packaging utilities — the list goes on. While exciting, this rate of change can be difficult to absorb, particularly when reliability, continuity, compliance, and security are more important than adopting bleeding edge technologies.

At the same time, leaders don’t want to be left behind or outplayed by their competitors. What, when, where and how to adopt has become a common dialogue between vendors and enterprise customers.

As vendors fight for their share of the cloud market, new strategies have emerged around complete platform control, deeply proprietary services, nested pricing models, and numerous methods to attract, ingest, process and retain applications and data. The typical vendor cloud strategy is becoming vertically integrated. Vendors want to capture as much of the stack as possible — SaaS, PaaS, IaaS, and the apps and data across, and ideally in a subscription-based model where the more you use, the more you pay.

Building a True Utility

These fundamental issues have kept enterprise cloud adoption in the low percentage points. But increasingly, enterprises are selecting and implementing cloud technologies that best fits their business needs — sometimes behind their firewall, sometimes hosted by others. Much like the evolution of other utilities, we are seeing a shift from a monolithic vendor-controlled world to a world where users decide their consumption patterns based on the needs of their business. We call this approach hybrid cloud computing and it is symbolic of the IT transformation we’re seeing today. The benefits of the cloud –speed, flexibility, self-service, on-demand consumption — are no longer exclusive to public clouds and are becoming borderless and universally accessible. Terms such as ‘public’ or ‘private’ cloud are losing their meaning as enterprises are building cloud into all of their IT plans. These hybrid models leverage their existing IT systems in new ways, and deliver substantial business value through self-service and automation on both the new and traditional apps and IT services.

Hybrid cloud is commonly implemented at the application level, either in the application platform — which offers flexibility in the choice of cloud infrastructure and delivery model — or in the application itself, designed to work across multiple cloud environments through APIs, app and data integration services, and networking methods to connect diverse systems. Cloud native application platforms are designed to abstract a lot of the infrastructure underneath, while aligning to a vital level in the business value-chain: end-user applications. And as enterprises move towards a cloud native application development and a DevOps culture, building hybrid cloud automation across the entire application lifecycle will be a critical step to provide the speed and independence that developers and business units are seeking.

Using the Past to Inform the Future

In the last decade, static web pages gave way to dynamic applications composed of a wide variety of web services, both in-house and from third parties. We call this the “Web 2.0” era, it was characterized by hundreds of third-party services that accelerated innovation by allowing application developers to focus on their unique value, while standing on the shoulders of others for more generic functions. Think of travel services like Kayak, or real estate experiences like Zillow, or Twitter trend aggregators like Trendsmap. All of them fuse a broad set of Web services to create something unique.

Our current cloud era is evolving in a similar manner, with application developers using distinct services from multiple clouds to create entirely new software. These dynamically composed applications often have components running across dozens of disparate infrastructures as cloud computing becomes democratized.

Just as we no longer think of the Web like we did in the 90s, we also don’t think of microcomputing, client-server, or even service-oriented architectures like we used to. Cloud computing is evolving to become more pervasive and practical for the enterprise than ever before. It’s like the HPC evolution, where we moved from proprietary, single-system supercomputers to distributed x86-based clusters coordinated and scaled with open source software. And also like the Web evolution, where we shifted from using one web browser and proprietary plug-ins for visiting web sites to an API-based, interoperable Web connected through nearly every device, cloud computing is unfolding from a singular model to a distributed, hybrid world where these capabilities are available to all users and builders, and to all types of businesses and governments — on their terms.

While many vendors are moving to verticalize and drive more platform control points, at Hewlett Packard Enterprise we believe the past will show that a horizontal, multi-cloud strategy is ultimately what will prevail. We believe that the future of cloud computing will be based on open technologies, innovative infrastructure, software and services that allow customers to evolve to a consumption model that makes financial and business sense for them.

In an accelerating present, the strategy we are executing is designed to match our customers reality as they transform to true digital businesses. The gradual but inexorable evolution of hybrid cloud computing for the enterprise makes today an exciting time for the entire IT ecosystem, presenting both opportunities and challenges for developers, IT, and business leaders aiming to keep up with the rate of change.

After all, if technology evolutions have taught us anything, it’s that history matters. Path dependence exists whether we like it or not and businesses must adapt, or risk being left behind. The Cloud — and the tools, methods and cultures that support it — won’t wait for the laggards to catch up.

Bill Hilf (@billhilf)