One of the biggest IT initiatives underway in recent times is migrating to a cloud-based delivery and consumption model. While public cloud adoption has been impressive and there are merits to having a private cloud, it is the hybrid approach that is appealing to many enterprises. Hybrid clouds promise the “best of both worlds,” i.e., the speed and scale of public clouds with the visibility and control of on-premises IT-owned private cloud deployments. They also give flexibility for enterprises to take a more measured approach to cloud adoption.
It is therefore no surprise that over 70 percent of enterprises polled were looking to deploy hybrid clouds in a survey by Rightscale’s State of the Cloud Report, while 18 percent were considering public-cloud-only options and an even smaller six percent were into private cloud only.
While the promise of hybrid clouds is no doubt appealing, it is also a complex problem to solve. Some call it the “wild west,” as there isn’t a set of consistent standards between private and public clouds or between the various public clouds. The operational model disparities and implementation challenges of VMware, OpenStack, AWS, Azure or Google Cloud Platform can all be quite different and necessitate a significant learning curve. For those adopting cloud-centric DevOps, the architectural changes, fragmented toolsets as well as fuzzy cost structures can be overwhelming.
Low integration between various clouds places the burden on system integrators and expensive professionals. The recent announcement of VMware and AWS integration is a welcome step in the right direction, but is by no means the nirvana that most enterprises are looking for.
Despite jumping through hoops, there are still no absolute guarantees on application performance, latency, security and even availability. Consequently, most enterprises are still taking baby steps with true hybrid clouds. This can be seen from the nature of hybrid cloud workloads in the industry that include backup and data archival, storage expansion, disaster recovery, dev/test and capacity augmentation. These are considered lower risk and not mission-critical workloads.
For most enterprises, all these deployment issues cause speed bumps in the move to both public and hybrid clouds. Even for simple workloads or for application “lift-n-shift,” these issues can delay adoption as development and test teams take a measured approach to moving workloads to minimize risk.
Over the past few years, I have come across several enterprises that planned a move to the public cloud or hybrid cloud and grossly underestimated the complexity of doing that; they frequently ran into issues that lengthened the timeline of “going live.”
This is where cloud sandboxes can help – particularly hybrid cloud sandboxes. The very nature of a hybrid cloud sandbox allows for replicating production environments including a full-stack on-premises environment (physical infrastructure, VMs, data, applications, APIs, test and performance tools, etc.) with the appropriate stack on any other cloud. Building these two disparate environments and running tests can ensure that any potential issues are caught early on with a greater degree of accuracy, thereby facilitating quicker adoption of public clouds or a well-orchestrated hybrid cloud deployment.
Leveraging sandboxes with cloud-centric DevOps automation can also de-risk cloud adoption in a cost-effective manner. Cloud sandboxes also allow for creating blueprints, establishing self-service workflows and letting IT operational teams establish governance mechanisms to get business insights with visibility into usage and metrics, minimizing impacts of shadow IT.
As the above illustration shows, an environment that has a VMware vCenter-based private cloud could continue to be used for traditional applications or in situations where data needs to be on premises for regulation or compliance reasons, while new cloud-native applications could be spawned in a public cloud. The hybrid cloud sandbox could also model secure connectivity based on existing VPN or other security infrastructure is set up for a realistic assessment based on a real-world production deployment.
Enterprises considering VMware-AWS, Azure or GCP-based hybrid clouds, or evaluating a move to OpenStack-based private clouds from VMware-based deployments, or potentially even moving from one public cloud vendor to another, can all benefit from leveraging cloud sandboxes and leveraging them as a stepping stone for hybrid and multi-cloud deployments.
If pre-production deployments in the dev/test stages holistically mirror the production environments, it goes a long way towards de-risking actual production deployments. Traditional on-premises vendors as well as public cloud vendors could also leverage hybrid sandboxes to smooth the experience for enterprise customers.
Shashi Kiran is currently the CMO at Quali. He has 20+ years of experience in the hi-tech information industry in areas of data center and cloud computing, security, internetworking, software and analytics for enterprise and service providers. He spent nine years at Cisco, where he was head of worldwide marketing for data center and cloud networking portfolio, and was most recently CMO at March capital-backed startup Appcito, acquired by A10 networks. Shashi can be reached at LinkedIn.