Achieve the Benefits of Cloud With Hybrid Cloud Maturity Model
In this article
Cloud computing provides organizations with flexibility and agility, allowing you to develop and deploy workloads quickly. A hybrid cloud approach offers you the ability to deploy workloads and data where they maximize value, meet requirements and solve business problems. It has the potential to be the best of both worlds — the cost optimization capabilities, agility, flexibility, scalability and self-service/on-demand capabilities of the public cloud, along with the control, compliance, security and performance of the private cloud.
However, many organizations do not experience the full benefits of hybrid cloud because they skip the necessary strategic work upfront. Organizations spent more than they needed to and achieved worse results than they expected by trying to boil the ocean and jumping into the cloud whole-hog without laying the necessary groundwork. Instant gratification simply will not get you where you need to go in the long run.
According to NIST, hybrid cloud is "a composition of two or more distinct cloud infrastructures (private, community or public) that remain unique entities, but are bound together by standardized or proprietary technology that enables data and application portability (e.g. cloud bursting for load balancing between clouds)."
In short, hybrid cloud is two or more of the other cloud models with a solution riding on top to enable application and data mobility between clouds. It is at its core an enabler of exit strategy. As an example, say I have VMware Cloud on AWS, Google Cloud VMware Engine and VMware Cloud Foundation on-premises. I can then move applications seamlessly between all three of those clouds using a solution like VMware HCX (a migration and internetworking tool), usually with little or no interruption to application service.
The true nirvana of hybrid cloud is the ability for applications to move dynamically and automatically at any given time to best achieve business objectives. Conceptually, this would be enabled via solutions that allow managing hybrid cloud through a single pane of glass — which is currently impossible for most, if not all organizations for many reasons, including the complexity of hybrid cloud. The single pane of glass management will continue to be a pipe dream for the foreseeable future because the software companies that develop such solutions don't control all the cloud platforms and, if they are a cloud provider themselves, lag behind the feature development of their own. This leads to cloud management solutions that are okay with one or two clouds while being really bad or even completely incompatible with all the others. Third-party vendors have offered unified visibility software, as well, but these solutions are also often lacking because vendors don't control all the cloud platform APIs and organizations don't want to pay for these tools. After all, day two operations usually gets the short straw when project budgets bloat and need to be cut back.
Although the ultimate dream of hybrid cloud is definitely a Herculean task, organizations can get increasingly closer to the end goal through continuous improvement following a logical, comprehensive approach to gaining cloud maturity.
The hybrid cloud maturity model helps organizations achieve the benefits of cloud in the most effective and efficient way.
Public cloud providers realized early on that you can't build ad hoc infrastructure and reliably know costs or effectively build automation. Instead, they put all the intelligence in the software and approached hardware as a commodity, building on open standards and using common features that all vendors typically offer. By staying away from the "special sauce" of proprietary things, they built the next gen data center so that it was standards-based, consistent and easy to automate. In these software-defined data centers (SDDC), they extracted cool features in the software service layer.
Likewise, the first step any organization should take on their cloud journey is to look at their islands of infrastructure, their various independent teams for unique cloud infrastructures, and traditional silos for network, storage, compute and security, to determine if it makes business sense to build (or acquire) a software-defined platform that can singularly handle a large chunk of the organization's application needs. This entails combining those traditional IT infrastructure teams, integrating with software APIs and relying on intelligence in the software service layer to provide most of the necessary features. If you skip this step, your cloud program will likely never reach its full potential.
There are two extremes as to how organizations may approach this step.
- Some organizations will prefer to "roll their own" private cloud through a piecemeal approach — potentially including open source and/or internally developed ones — to provide the relevant services, integrations and automations. These organizations will find themselves spending a significant amount on personnel with the highly sought-after talent necessary to achieve this end.
- Other organizations will prefer to purchase their private cloud, choosing off-the-shelf components designed for tight integration with one another and sharing support by a single vendor, such as VMware Cloud Foundation. In this case, organizations will find themselves spending a significant amount on products while tapping a wider — though still highly sought after — talent pool to administer and maintain everything.
Most organizations will fall somewhere in between these two extremes as their day-to-day operations are heavily influenced by challenges such as industry compliance, security requirements, technical debt, existing tools and skillsets, and specialized use cases.
When it comes to the private cloud, it requires its own maturity model. During this step, a lot of organizations either are or will get hung up here. Additionally, decisions made during private cloud selection may hinder or even impede the organization's ability to continue progressing into a hybrid cloud model. Choose incorrectly, and the organization may find the ability to move applications and data between clouds without re-platforming to be lacking.
The hybrid cloud model enters the picture during this phase. Feature and service abstraction through intelligent software can extend to multiple clouds, enabling application and data mobility.
At this level, you begin to define and adopt cloud native networking and security best practices. This is done through a combination of technology (e.g., virtualizing the functions and roles traditionally fulfilled by physical appliances) and people/process (e.g., combining your cloud infrastructure and security teams, or at least making them work together, to architect and document what "good" will look like).
At this point, you begin to deploy your hybrid cloud platform with the SDDC and enterprise service management systems. You begin to use cloud deployment selection best practices, migrate applications and start hybrid cloud data protection.
Also, at this level, you gain insights into important areas, including:
- True costs of running your critical applications in a given cloud environment.
- Application experience, usually related to network adjacency to users, data and other application tiers.
Public clouds are "closer" to remote users since they have data centers all over the world, so they can provide better experiences in terms of lower latency. This was especially helpful for organizations who could move services such as virtual desktop infrastructure (VDI) to the public cloud during the pandemic, thereby getting closer to remote users wherever they were. Improving geographic distribution of services is also important for companies pushing to mobile applications who want to provide the best possible experience for their customers. At the same time, there are often compliance and security concerns with data or even critical applications that require more care and feeding than the public clouds can provide. In these situations, the private cloud will be more advantageous.
You can use the insights gained in level 4 to move applications and data around to the best places for them to achieve your business objectives. In this final phase, organizations will attempt to employ automation so that those applications and data move on their own based on quality data metrics and perform backend testing to ensure the migration was successful. But achieving this level of maturity requires more than just automation skills. It also requires a deep understanding of multi-tiered application dependencies such as data, other tiers in the same stack, other application stacks entirely, and even enterprise-wide services such as SSO (single sign-on) and DNS (domain name system). These are the exact same insights that often slow down or even break today's cloud migrations, and understanding them is also critical to getting the most value from your hybrid cloud.
Organizations that achieve this highest level of hybrid cloud maturity will be able to monitor and optimize applications at or near real time. This includes the data metrics necessary for governance, security, compliance, performance and alerting. It should also include visibility into costs (through monitoring and reporting) which is necessary for both application rationalization as well as the placement recommendations for new and existing deployments.
In the real world, few organizations will follow this approach closely. Instead, they will skip over steps or jump into something prematurely, often to their own detriment later. The good news is that you can use the maturity model as a tool to see where your organization currently lands, as well as identify the steps that may have been skipped and the steps that are still remaining along the path to realizing the most benefit from your hybrid cloud.
First off, accept the fact that you can't get there without investing resources into IT. The bare truth is that cloud is cheaper — except for all the work it takes to get there. On the other hand, all that work will make your business better in the long run and needs to be done anyway. You can think of it like an advertisement for a "free house" where, while the house itself is free, the fine print on the offer notes that you'll need to move the house yourself to your own land.
The best thing to do is to start with new applications and move from there. You may also consider containerizing the new applications since container platforms by their very nature fit the hybrid cloud definition. Set a line in the sand and move forward with this greenfield approach. You could also take a brownfield-ish approach by adopting and migrating to a new model as part of a life cycle refresh of legacy infrastructure. Or, maybe it's a combination of both approaches. In either scenario, you would build to a standard and progress forward a little bit at a time. Remember that you can't eat the whale in a single bite; you have to start small and grow.
And, it may be a slow and tedious process to get there. Find opportunities where tools and technologies are familiar, such as extending a software-defined data center solution that allows you to capitalize on existing expertise and operational tools. This will get around re-platforming speed bumps, such as VMware virtual machines not being compatible with AWS and AWS virtual machines not being compatible with VMware. You may also consider refactoring a legacy monolithic application into microservices, which is typically done in containers. Remember that containers by their very nature are able to run in almost any environment.
In summary, you have four main pathways to hybridity. The options include:
- Share software-defined infrastructure between clouds (e.g., AWS sends their infrastructure on-premises or you run VMware on AWS).
- Move to containers, which are natively portable.
- Adopt migration tools that translate between different clouds.
- Use third-party solutions that make native cloud services look like and be compatible with known systems (e.g., NetApp Cloud Volumes ONTAP for storage or Cisco Cloud ACI for network and security).