In the beginning, the only “real” cloud was public cloud. There were those that thought that it wasn’t a fit for the enterprise, or perhaps would only be used for test/dev environments — and there were plenty of historical failures for solutions before public cloud to bolster these assertions — today we know that public cloud is a significant driver of innovation and change in the IT landscape. Private cloud was derided by the public cloud crowd as nothing more than “legacy mindset” or a desperate attempt by box sellers to stay relevant by co-opting the cloud term. Wikibon put forth the term “True Private Cloud” in 2015 to show that cloud is as much an operational model as it is a technology (and private cloud needs to be more than virtualization+). Back in the early 201x’s, hybrid cloud was often discussed as a “halfway house” on the journey to cloud, or part of a “bi-modal” strategy. For most enterprise companies today, hybrid cloud (having both public and private cloud) and multicloud (having multiple clouds, a superset of hybrid cloud) is just as much a reality as multi-vendor was in the past. The premise of Wikibon’s research on hybrid cloud is that there are many reasons why hybrid and multicloud solutions are needed, and even preferable.
At the bottom of the article is a 11 minute video playlist with the CEOs of AWS, VMware, and Nutanix, one of the creators of Kubernetes, Red Hat, Cisco, and IBM executives discussing hybrid cloud.
Why Hybrid
There are many reasons why a single solution would be easiest – a single skill set, leverage on pricing, and simplified operations and interoperability are just the beginning. There is a general consensus on some of the reasons why hybrid cloud is needed – the CEOs of Nutanix, VMware, and AWS have all talked publicly about the “3 laws of cloud” (Pat Gelsinger shares these 3 laws in the first video clip below, I’ve heard almost identical descriptions in keynotes from Dheeraj Pandey and Andy Jassy):
- The laws of physics – latency and the speed of light are limiting factors for the distance between my data and compute for many applications.
- The laws of economics – bandwidth and data movement costs money (reminder: 5G will not make the laws of physics and economics disappear)
- The laws of the land – governance, GDPR, etc. Edge and IoT will make this more prominent.
Public cloud solutions are often described as frictionless, and good private cloud solutions focus on simplicity; it is early days in hybrid cloud, which today is neither of those terms. Hybrid cloud today is less of a product or solution than it is a combination of things. Wikibon’s David Floyer put forth a Hybrid Cloud Taxonomy in order to understand the major differences between the myriad offerings on the market today, and where we believe they should be heading. The taxonomy is a broad spectrum that digs into the nuances and complexities of hybrid cloud, multicloud, and edge computing (watch out for terminology that confuses or conflates edge with data centers). The more important thing to look is not what term is being used, but how application architecture and the role of data is impacted by the maturation of these solutions. Applications have moved from monolithic to microservices, as discussed by IBM Fellow Bala Rajaraman, there are dimensions of complexity in programming models and integration, data movement and governance, and security.
Key Players
Since there is not general consensus around hybrid cloud solutions, and from a taxonomy standpoint, our position is that this is a number of markets, there is not a single market leader (so expect many vendors to plant a flag claiming leadership on the hill that they have climbed). In public cloud, each major player has a different positioning, and while AWS has leadership in IaaS, Microsoft is a strong second with many business productivity SaaS offerings, Google has a strong play with data services. These leaders in public cloud are all of major importance as part of examining hybrid cloud. So do the leaders in data center and private cloud – VMware and Dell, Microsoft (again), Oracle, IBM, Cisco, HPE, and many others have large install bases that are in the process of creating private clouds, modernizing applications, and highlighting the value of data with AI. The relationship between the big 3 public cloud providers and other vendors changes at a rapid pace. AWS has a sprawling ecosystem and the AWS Marketplace is likely the most important for software today. The Microsoft of today is a kinder and gentler partner than the one that many of us remember from the days when Windows wielded its dominance to crush competition. Whole ecosystems have been created from technology that sprung from Google papers (see Hadoop, Kubernetes, etc).
Considerations and Opportunities
On Kubernetes, co-creator Joe Beda said “we never wanted to pretend that Kubernetes was going to be this magic layer that was going to make differences between different environments disappear. What we did want to do, though, was actually find the commonalities and minimize the extra differences that didn’t need to be there.”
Management of heterogeneous environments has been something that IT has tried to address for decades – the fabled single pane of glass is typically more pain for the administrators who have to fight with it. There is hope that modern architectures can leverage APIs and lead towards a more manageable environment; as Cisco’s Ronnie Ray said, “software can have a unifying effect” towards platforms and marketplaces. Speaking of Cisco, networking between environments is one of the planes of the taxonomy. SD-WAN is a fast growing solution due to its enablement of hybrid, multicloud, and edge solutions (see more in my article Cisco’s Software Bridge to Multicloud). While SaaS offerings are part of the cloud landscape, there are a number of services that should be considered in the context of hybrid cloud, most notably security and data protection. The reason to call these out specifically is that they are business critical services that if not considered holistically across all environments, can put companies at risk to data loss or security vulnerabilities. The IBM $34B acquisition of Red Hat highlights the opportunity in hybrid and multicloud; Red Hat Linux (RHEL) and OpenShift are supported across most public and private cloud environments, proving the consistency across environments that customers are looking for in this complex landscape.
Action Item
Data and the application portfolio that interacts with data are the starting points to determining the services needed, and therefore the architectures that can deliver on business requirements. The balance between standardization and flexibility is more challenging than ever. Trusted partners that can educate and iterate in this fast changing world are critical.
Highlights included below: Pat Gelsinger, CEO of VMware sharing the 3 laws of cloud; Dheeraj Pandey, CEO of Nutanix on the honest solutions needed in hybrid environments; Andy Jassy, CEO of AWS discussing why customers need on-premises solutions; Joe Beda on what Kubernetes is and isn’t; Red Hat’s Ashesh Badani on customer reality of hybrid cloud; Cisco’s Lew Tucker on the network plane of clouds; IBM’s Bala Rajaraman on the changing application architectures in cloud environments.
