Kubernetes & Containers at Scale are Difficult
Organizations are beginning to leverage containers for digital acceleration and to simplify their application development pipelines. As environments scale and grow, managing and enabling a large scale containerized infrastructure becomes increasingly difficult.
Enterprise Architecture is the most difficult thing to get right in a cloud-native world. While organizations plan to deploy digital applications or rehost legacy applications in a hybrid cloud, several complexities present themselves. However, the solutions to these complex undertakings don’t result in a similar set of customer rewards.
This discussion has particular relevance to Kubernetes, which has emerged as the defacto infrastructure and application standard in the cloud. However, Kubernetes is a complex technology and can get complicated to deploy in a large-scale hybrid environment.
The three fundamental & key enterprise use cases in the context of containers remain as follows:
1. Manage and monitor Hybrid Clouds: Manage environments that consist of on-premises, public and private clouds.
2. Cloud-native application development ensures consistency: Leveraging containers and Kubernetes ensures consistency in application development and deployment.
3. Compute resources at the edge: Organizations are finding more and more need for resources to be closer to where information is generated for faster analysis.
The Two Broad Areas of Container Enablement
Let us break these challenges down into two broad areas. When enterprises adopt the cloud there are two broad categories of challenges –
- Infrastructure Management – Deployment, Monitoring, and Upgrades
- Product Development – Architecture & Development of solutions
Of these two categories of challenges, much higher business value is related to the standardization, maturity, and repeatability on product development as opposed to investment infrastructure management. Innovation resides here as well as business optimization. Further, product development brings in allied disciplines – business processes, integration with existing applications, partner relationships and most importantly new age technology. Infrastructure Management merely adds value to product development but does not standalone in and of itself. Mature enterprises understand this and focus on customer relationships, new product development and bringing existing products into the digital stream. This is also seen in the industry where enterprises are focusing on rationalization and consolidation as ways to cut down infrastructure spending by focusing on a private cloud or investing in the public cloud.
As depicted in the illustration above, core Kubernetes features that provide value to an application are depicted on the left while supporting infrastructure elements are depicted to the right as – Enterprise “Must Haves”.
Core Kubernetes – The Kubernetes project provides fundamental constructs & API to think of an application in terms of containers. It also provides the ability to deploy these across multiple nodes or server hosts. Once deployed, it provides the orchestration, management to scale and then manage the health of these containers.
Enterprise “Must Haves” -For organizations to get value from containers, fast, with container orchestration delivered as a service, there are additional capabilities that they must provide for your deployments to function in a multi-tenant manner. These range from Self Service to CI/CD pipelines to Serverless functions to Site Reliability & it’s automation.
Whatever the business project, it is a sure bet that containers will spread across your enterprise in the months and years to come. There will not only be traditional servers to manage but also complex networking, storage, and security implications. Handling all these requirements implies hiring expensive cloud admins as well as writing a high degree of automation to handle repeatable tasks. The point is well made that any error or sub-optimization in the infrastructure layer can escalate quickly to cause customer dissatisfaction.
Why the Managed Service Is a Critical Enabling Capability
A managed service provider such as Platform9 provides CNCF approved Kubernetes that can integrate well with all of the enterprise services on the right while helping operations teams with the vast majority of the heavy lifting associated with containerized infrastructure such as –
Day 1 Operations – Self Service led installation, provisioning, configuration of networking, storing & tuning
Day 2 Operations – DevOps pipeline setup, cybersecurity based hardening, integration with DbaaS ( Database as a service), Monitoring etc
Day 2+ – Single Pane of glass for all kubernetes environments across a hybrid cloud, kubernetes upgrades, rollbacks, roll forwards etc
A poorly thought out and implemented container infrastructure can result in significant business risk in the following ways
- Inefficient & inflexible provisioning cycles
- Developer dissatisfaction
- Lack of visibility into the overall environment
- Lack of developer self service
- Cost overruns