The cloud has become an indispensable part of the internet as it facilitates many of our elementary consumer and business applications. As technology is progressing and consumers and businesses demand certain functional requirements, the cloud is undergoing some noticeable changes. Here we take a closer look at a few interesting changes and what the direct and indirect consequences are for applications and businesses.

Our observations

  • According to Synergy research, the cloud market is currently dominated by Amazon (33%), Microsoft (13%), IBM (8%), Google, (6%) and Alibaba (4%). When only looking at public cloud services (IaaS and PaaS) we see that Amazon, Microsoft, Google and Alibaba take the cake. For 2020, CITI expects AWS at $44 billion, Azure at $19 billion and Google Cloud Platform at $17 billion.
  • IBM acquired Red Hat by buying all shares at $190 resulting in a deal valued at around $34 billion. According to their own press release, IBM’s main focus is to become a hybrid-cloud provider that is able to bring portability and security to multiple clouds. This also seems to point towards a consolidation of the containerization industry.
  • As we described in our outlook for 2018 there is a growing need for multi-cloud and hybrid-cloud solutions as a consequence of enterprises that are seeking ways to escape vendor lock-in, gain flexibility and mitigate risks.
  • Interestingly, long-term largescale cloud storage is facing issues as it relies largely on a concentrating magnetic tape market which is now further being jeopardized by a legal battle between Fuji and Sony.
  • At the Akamai Edge EMEA Forum, Tom Leighton laid out the company strategy regarding edge computing in which they focus on 5 areas: media distribution, performance, security, internet of things and blockchain.
  • Enterprises are migrating their workloads from colocation data centers to the cloud. Consequently, colocation retail is slowing down while the wholesale of data centers is increasing. In response, data center companies like Equinix are betting on multi-cloud businesses to compensate for slowdown.

Connecting the dots

Even though cloud computing as we currently know it was introduced by Amazon’s AWS Elastic Compute Cloud product in 2006, the principle of virtualization and time-shared computing can be traced back to ARPANET and remote job entry (RJE) in the 1960s. Subsequently, we have gone from cluster-computing, to grid-computing to the multibillion dollar space in which the tech giants roam. With this development we can see that the internet has undergone immense centralization with the purpose of achieving scale, performance and cost efficiency.
However, now that the migration of large enterprises to the cloud is well underway we can see that the emphasis on cost and performance has gradually shifted to other mission critical requirements such as interoperability, data portability, anti-vendor lock-in and security. In order to achieve these goals, the cloud industry has put its focus on enabling enterprises to distribute their workload among multiple computing environments, whether it be on both private and public clouds (i.e. hybrid-clouds) or on multiple public clouds (i.e. multi-cloud). Thereby we actually see that virtualized computing is becoming more distributed.
An important enabling factor in creating these cloud-agnostic solutions is the use of containers. Containers allow for the

quick and easy deployment of discrete components of application logic that are able to run anywhere.  Its popularity can already be seen by the attraction of many companies that form around different parts of the containerization solution stack, whether it be container operating systems, container engines (e.g. Docker), container orchestration tools (e.g. Kubernetes) or application support services. Interestingly, many of these companies deploy open-source code in some way so that open standards can easily form for the purpose of interoperability.
Going forward, an important determinant for how the cloud will further evolve is the advent of edge computing. With the development of powerful specialized chips, AI systems and the rollout of the 5G network, we are able to develop low-latency smart applications ranging from swarming autonomous vehicles to virtual personal assistants. Hence, similarly to the advent of cloud computing, which was partly driven by the smartphone adoption, we will probably see a shift from cloud to edge computing driven by these types of applications. As such, there is a possibility that edge computing will draw away some of the workload from cloud services, potentially leading to a possible reshuffling of the pecking order within the virtualized computation space.

Implications

  • When container orchestration (e.g. Kubernetes) will become more efficient and smart, multi-cloud serverless systems could lead to price arbitrage as enterprises try to dynamically move workload based on CSPs that are the most price competitive. In contrast, there are also easier ways to achieve price efficiency by actually aiming for volume discounts at individual CSPs.
  • As the cloud space becomes more and more commoditized, CSPs will feel more pressure to develop integrated proprietary cloud services (e.g. blockchain, AI) to differentiate themselves from the competition.
  • As noted previously, China has been vocal regarding its ambitions to become dominant in each layer of the tech stack, whether it be in semiconductors, 5G, blockchain and AI. It can be expected that Alibaba and possibly others will try to compete with Western cloud services.
  • With the shift from cloud to edge, the value of the network could also shift away from cloud server software to specialized edge hardware (e.g. SoCs, ASICs, FPGAs) that are necessary for low-latency applications.