How to run high-performance Kubernetes apps in the multicloud
Cloud services will account for virtually all data center traffic within the next three years.
The evidence: Last year, a report from the Cisco Global Cloud Index predicted that 94 percent of workloads and instances will be processed by cloud data centers in 2021, while a 2017 International Data Corp. study showed that 87 percent of enterprises have deployed or plan to implement a hybrid private/public cloud model.
Going forward, most enterprise workloads will be processed in multicloud environments of growing complexity. More of those deployments will rely on hyperscale data centers that deploy compute, storage and other hardware resources on a new generation of hyperconverged infrastructure platforms. And much of it will run in containerized software environments that are orchestrated and managed in multicloud Kubernetes deployments.
Kubernetes is the foundation for portability, scaling and management of containerized application workloads across multiclouds. Running high-performance Kubernetes apps across heterogeneous public and private clouds will become critical for most enterprises in the coming decade. In addition, Kubernetes will drive agile orchestration and resilient load-balancing of application traffic across multiclouds.
For enterprises deploying containerized workloads across hybrid and multicloud environments, the journey to high-performance Kubernetes includes the following principal steps:
- Deploy hyperconverged infrastructure to simplify scaling of hardware platforms running Kubernetes in the multicloud: In their on-premises data centers, enterprises should adopt hyperconverged data center infrastructure. This simplifies scaling of on-premises compute, storage, memory and bandwidth resources for Kubernetes apps running in large, distributed clusters. The hyperconverged infrastructure components should natively support all container, hypervisor and serverless platforms that are used by developers to abstract their applications away from the underlying hardware. It should provide persistent storage for containers managed using the Kubernetes orchestration platform, enabling deployment of cloud-native apps inside data centers. To reduce business exposure to data loss in the multicloud, the infrastructure should also support storage replication, backups that are agentless and application-aware, and enforcement of stringent recovery point objectives.
- Implement a federated orchestration backplane for Kubernetes apps: Enterprises should build their multicloud application environments on federated Kubernetes deployments that interoperate seamlessly across public and private clouds. This should involve deploying container and data hub platforms that have been optimized for hybrid cloud deployments. These platforms should have been prevalidated to scale on the infrastructure platforms that users are running in their private clouds and providers in their public clouds. Through infrastructure-as-code tooling, information technology administrators should be able to automate programmable provisioning of compute, storage, bandwidth, memory and other hardware resources across public and private clouds. Developers should be able to deploy their containerized apps and constantly monitor those apps’ performance anywhere on the multicloud. They should be able to automate and scale deployment of containerized apps, balance workloads, enforce policy-driven security and detect and report on dynamic conditions at the application or cluster level.
- Centralize management, optimization and securing of Kubernetes apps, workloads and traffic: Developers should be able to easily manage, monitor, load-balance and secure containerized workloads throughout the multicloud. The tooling should make it easier for enterprise IT to develop containerized applications that span Kubernetes clusters in private data centers and on public clouds. It should enable applications to execute in the public clouds of the enterprise’s choice while retaining the ability to keep any or all workloads on-premises. It should simplify the provisioning and management of containers at scale across hybrid, mesh and other multicloud configurations. It should provide a unified repository for storing container images and prepackaged versions of the most commonly used software components across the multicloud. IT administrators should be able to sync up their on-premises and public cloud Kubernetes configurations so applications may be moved among them without requiring major changes and without developers needing to know the configurations of target environments.
Another key step in your enterprise multicloud journey is to attend Cisco Live, which is taking place Jan. 28 to Feb. 1 in Barcelona. Please tune into theCUBE for live interviews with Cisco executives, developers, partners and customers during Cisco Live.
A message from John Furrier, co-founder of SiliconANGLE:
Show your support for our mission by joining our Cube Club and Cube Event Community of experts. Join the community that includes Amazon Web Services and Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger and many more luminaries and experts.
We really want to hear from you, and we’re looking forward to seeing you at the event and in theCUBE Club.