The cloud-computing market faced some unique challenges in 2013, including the fact that moving clouds into production means new and difficult issues, many of which were not initially understood. Enterprises need to understand how to work properly with cloud-computing technology. In 2013 they had to learn how to design applications to take advantage of cloud-native features such as the ability to self- and auto-provision, as well as self- and auto-scale. Enterprises also learned to define strategic architectures that go beyond a single cloud deployment, and figured out an approach and management technology to make the cloud resources all march in the proper directions to provide automation in support of resources governance and DevOps.
There were some obvious bummers in the emerging cloud computing market. Many enterprises found out that cloud-based platforms may not be as cost-effective as we initially believed. This is due to unexpected operational costs, and migrating the wrong applications and data sets to the cloud. We’re now in the process of dialing these metrics back into the cloud/no-cloud decisions.
On the upside, while many found that cloud computing was not a fit, more found that cloud computing does indeed live up to most of the promises made in 2013. This includes the ability to operate core business applications at a fraction of the cost of traditional platforms, and the ability to finally bring business agility and time-to-market advantages to the enterprise because of how easy it is to provision and deploy public cloud resources.
What worked in 2013
The dominance of AWS became very clear in 2013. Most analysts put the company at roughly a $20 billion USD total valuation. As AWS continues to nail the market, that valuation seems conservative. AWS’s re:Invent conference held in November was packed, and provided enough data points that the company is indeed a vendor that can play in the enterprise, which many felt would be its Achilles’ heel. Thus, far AWS has not shown many weaknesses, and seems to be defining the standard for IaaS clouds, and that has the other IaaS providers worried.
OpenStack grew in 2013. This was clear with the release of Grizzly in March, and the release of Havana in November. OpenStack turned three years old in 2013. It has a lot of momentum to show, including adding pretty much everybody but AWS to its ranks, including HP, IBM, and now Oracle. While you would think that this “dream team” of technology powerhouses would be enough to push down AWS a few clicks, the growth of OpenStack has been more within the private cloud portion of the market, where AWS does not play.
The cloud proved itself green. The release of new information from Berkley Labs supported the assertion that cloud computing is truly energy efficient. As the cloud providers build data centers the size of blimp hangers, the greenness of the cloud was called into question. However, given the sharing and the consolidation effect that cloud brings, energy use should go down, relative to the number of users.
The cloud can be managed though the use of cloud management platforms. These technologies are not new, but they found a new purpose in 2013 as multicloud deployments became the way to go. These tools handle provisioning and deprovisioning of cloud resources, such as storage and compute services, as well as place policies around the use of those services — even the ability to monitor and charge back for those services. The ability to provide a “single pane of glass” approach to cloud management removes the enterprises from having to deal with the underlying complexities, considering that cloud-based deployments are truly complex distributed systems at their essence.
The cloud proved itself. In other words, most deployments worked as advertised. In 2011 and 2012 we were largely at the POC stage with cloud computing as enterprises dipped a toe in. In 2013 we saw a few massive deployments to the cloud, with mission-critical applications running in production within the major cloud-computing providers. The world did not come to an end and the cloud did not drive massive layoffs, as many predicted. Some of the lessons learned include the fact that applications need some modification to take advantage of cloud-native features, such as provisioning services. The data typically should reside with the applications, when dealing with the public clouds, for performance and security purposes. Security and governance required some advanced planning.
What did not work in 2013
The dominance of AWS showed the strength of the cloud computing market but also represented a single cloud provider that got very big and powerful very fast. To many in the market, that brings to light some negative consequences. While somebody had to be in first place, the gap that AWS is building between itself and its closest competition widened in 2013, with Microsoft or Google being in second place, depending upon whose market data you read. AWS is already setting market prices for cloud services, driving many IaaS startups to drop well below profitability. The larger players that are spending billions to get deeper in the-cloud computing market are perplexed by the rise of AWS, and are not sure how to respond. AWS does not carry the legacy baggage of an IBM, Oracle, or HP, and thus can be more nimble and innovative. The downside for the user is that we could end up with a de-facto single solution for public cloud IaaS, which could make the market much more homogeneous much quicker than we had thought. Think Microsoft in the 1990s or IBM in the 1980s.
The NSA rained on our cloud parade, with some analysts firms predicting the PRISM scandal could cost cloud companies as much as $180 billion. This also placed cloud computing on center stage for the press, and the general public questioned the wisdom of the cloud, given proof that the government could cull through your personal and business data. While this did have a sobering effect on the emerging cloud computing space, the big pushback that everyone expected did not really happen. Or, if it did, the effect cannot yet be measured. While companies outside of the U.S. are indeed concerned, companies in the U.S. have dialed this fear into the mix along with the other fears around cloud usage. We have moved on, it seems.
Cloud companies went away, such as cloud storage providers MegaCloud and Nirvanix. This is perhaps the start of the dark side of cloud computing, where those holding your applications and data are here today and gone tomorrow. You’re left out in the cold, and the use of cloud-based resources has actually caused you to take a step backward. In these events, and likely other events that will occur in 2014, we’re learning to dial in viability and stability when we select cloud providers. This could mean another reason avoid smaller providers, and thus we have a narrower field of cloud-computing competitors.
Trends and technologies to watch
The continued growth of CMP and other management platforms continues to drive the growth of cloud computing for larger enterprises. The ability to mix and match cloud-computing providers and other cloud-based technologies is a key enabler for enterprises that seek to leverage best-of-breed in the world of the cloud, both private and public. Cloud computing is now a team sport. Focus will move from simple operations — e.g., the ability to add, remove, monitor, and manage cloud resources — to DevOps that includes support for continuous integration and continuous development. Count on more consolidation in this space as many of the smaller players, such as Gravitant and RightScale, follow ServiceMesh’s lead and move to larger players.
Enterprises will move to a security model that is a better fit for the distributed nature of cloud computing. This means security solutions that support identity-based approaches, such as Ping Identity and others, will be in increasing demand. Cloud providers will quickly learn how to provide the right guidance, perhaps building identity-based security systems in their cloud that can be had on-demand.
PaaS has found its place in DevOps. While PaaS has always been a part of cloud computing, most enterprises have yet to find it a proper place in their cloud computing strategies, and resulting technology stacks. Merging PaaS capabilities with DevOps automation will be a pairing that many enterprises will find valuable. Popular PaaS providers have been AWS, Microsoft, and Google, among others. PaaS’ place is to provide automation around the delivery of business applications, as well as the ability to continually update the capabilities of that solution to meet the needs of the business with reduced latency. The focus on automation, as well as continuous delivery and continuous integration, brings a new value to the use of PaaS providers in a space that has been somewhat confusing and thus bypassed by most enterprises.
AWS becomes the standard bearer, as discussed above. AWS is now in the same place as IBM, Microsoft, and Oracle in defining the standard for its technology space. It will continue to lead the trends in pricing, technology, and use cases, and will create a growing ecosystem around AWS that will drive the growth of other technology providers. This dominance will call many things into question, frustrate the existing technology powers-that-be, and provide enterprises with very simple choices when it comes to selecting a strategic IaaS player. The positive or negative effects are yet to be determined.
The trend is your friend when considering cloud computing. The big players will become bigger, cloud management and governance will become more useful, and security will become more important. The year 2013 was all about getting serious with the use of cloud-based platforms, and most enterprises found success in the cloud. While the cloud is not to be used for everything, it should be a core consideration when an enterprise needs to improve time-to-market for business-critical applications, improve business agility, and reduce costs.
The challenges moving forward will be those of planning and architecture to deal with the complexity that cloud computing brings, including the movement to multicloud, which can provide more value through heterogeneity, or by providing choice of cloud resources to leverage. Most enterprises have a good handle on cloud computing, at least conceptually, but most have just begun to work with this technology. The learning curve will be steep, and a few failures are inevitable.