Why Object Storage Is Key in the Cloud Operating Model
No matter where your enterprise cloud exists, one thing is almost certain – that Kubernetes, containers and object storage are also there.
The cloud now defines the enterprise technology landscape. It is an operating model, and it is massive. But, thinking of the public cloud as “the cloud” will bring limitations. Anand Babu Periasamy, CEO and co-founder of MinIO, discusses the attributes of the cloud operating model and delves deeper into what makes object storage primary when it comes to this model.
The enterprise technology landscape is driven by the cloud. It is an operating model – not a physical location. It is easy to equate the public cloud with “the cloud,” and it is massive – with Gartner estimating the public cloud to be a $494B a year market. But thinking of the public cloud as “the cloud” will limit thinking, reduce flexibility and introduce lock-in.
That is why established enterprises and emerging startups are changing how they think about the cloud. This is not to suggest that the public cloud is not revolutionary – it has changed how enterprises analyze, store and visualize data. The public cloud is where most organizations learn “the way of the cloud,” and it offers the allure of on-demand infrastructure, a bevy of services and minimal friction.
Still, over the long haul, the principles of the cloud matter more than the location.
The cloud can exist at the edge, in AWS/Azure/Google or on-prem. No matter where your enterprise cloud exists, one thing is almost certain – that Kubernetes, containers and object storage are also there. That is because Kubernetes and object storage are the foundations of multi-cloud infrastructure.
See More: 6 Top Considerations as You Evolve Your Cloud Data Management Strategy
The Cloud Operating Model and Object Storage
The term Cloud Operating Model should be credited to A16Z’s Martin Cassado. While he doesn’t use the term in his highly controversial and superbly researched blog entitled The Cost of Cloud, a Trillion Dollar Paradox (full credit also to co-author Sarah Wang) – he used it extensively in his subsequent interviews and podcasts about the article. The concept echoes former VMware CEO Paul Maritz’s comment, “Cloud is about how you do computing, not where you do computing.”
The cloud operating model speaks to engineering first principles, DNA and mindset. What, however, are the attributes of the cloud operating model?
- The cloud starts with containerization: Indeed, 92% of enterprises use them today. Containers encapsulate a lightweight, immutable runtime environment and all its components, including the dependencies: libraries, runtime, code, binaries, and configuration files. They visualize the operating system (OS) and can run multiple workloads on a single OS instance.
- The cloud runs on Kubernetes: Kubernetes orchestrates deployment, scaling, and management of the containers above but does much more. Kubernetes makes infrastructure look like code, delivering full-scale automation to both stateful and stateless components of the software stack.
- The cloud is software-defined: You cannot take hardware appliances to the cloud. Beefy virtual machines are an anti-pattern in the cloud. The idiomatic way is to containerize your software stack and use Kubernetes to orchestrate them across public and private clouds uniformly. Software-defined (or cloud-native) approaches provide greater flexibility in economics, capacity and performance.
- The cloud is automated: The scale and complexity of the software stack preclude humans from actively “managing” the software stack. Resilience (a nice word for failure) is handled gracefully and seamlessly and, when architected correctly, should not require humans. The same automation mantra applies to configuration management and security.
- API is driven as opposed to interactive GUIs: The cloud operating model will require that APIs are used to make everything a service. Large monolithic systems are broken down into modular services (stateless microservices when possible), and they talk to each other via APIs. The API-driven approach allows automation, resiliency and significantly lower cost of operations at scale. The cloud operating model also strongly suggests the “loose coupling” approach to APIs such that lock-in is avoided.
- Elasticity is key: The cloud operating model is about scale. It stands to reason that scaling seamlessly and without constraint are requirements. The bigger the scale, the more technologies will be eliminated from consideration. What’s left is cloud scale.
- The cloud is multi-tenant: Multi-tenancy allows multiple customers to share the same infrastructure and, when implemented correctly, can reduce operational overhead, decrease costs and reduce complexity, especially at scale. However, it also requires strict resource isolation so multiple users can access compute or storage resources without impacting other users.
- Running in user space: For an object storage solution to be cloud-native, it has to run entirely in the user space with no kernel dependencies. This is not how most object storage systems have been built, particularly the hardware appliances. Nonetheless, if you want to containerize your storage and deploy it on any Kubernetes cluster, you must abide by these constraints. By definition, this means solutions that require kernel patching or have specialized hardware won’t be cloud native.
- Built with microservices: The cloud-native way is to break monolithic systems into smaller services that can be updated and deployed easily and frequently. If those services can be stateless – all the better.
- A difference in culture: The devops culture fundamentally differs from the enterprise IT mindset. It is difficult to change and often requires new people in new roles. If the enterprise does not plan for this cultural change, it will only accrue a small fraction of the benefits associated with the cloud operating model. Developers are the engine of value creation, and the organization needs to reflect that in its structure.
Here is a simple litmus test. If your entire software stack can be defined in a Kubernetes YAML that can be deployed on any new infrastructure, public or private cloud, several times a day, you are working in the cloud operating model. You will not pass the test if there are proprietary services, hardware appliances, or bare metal software dependencies.
How are you leveraging object storage is your cloud operating model? Share with us on Facebook, Twitter, and LinkedIn. We’d love to hear from you!
MORE ON STORAGE:
