Containers and Kubernetes allows enterprises to build applications that are portable across clouds, and provides seamless mobility between infrastructure pools. However, when an application instance has state, this flexibility disappears, and they must rely on mechanisms based on data migration to deal with the state. Data Teleport™ allows stateful applications instances in Kubernetes to be instantly copied or moved between datacenters and clouds by providing instant mobility for persistent volumes.
Enterprise Clouds and Kubernetes
Enterprises are transforming their IT from traditional siloed deployments to clouds that span across multiple data centers, hybrid clouds, and multi-clouds. This transition is motivated by the need to make IT more effective in serving the businesses goals – reducing costs, delivering better service to end users and increasing agility while navigating security, vendor lock-in, and regulatory and compliance requirements. The ultimate goal is a unified pool of IT resources and services across multiple infrastructure silos and providers – an Enterprise cloud that is distributed and would allow seamless mobility of data and applications to enable business to deliver services to end users when and where they are required with minimal infrastructure and operational costs. The rapid adoption of Kubernetes enables customers to move closer to the desired future of a distributed enterprise cloud.
Between 2018 and 2020, Kubernetes use has increased from 28% to 48%, according to a study conducted by VMware. In addition to allowing IT applications to run anywhere, Kubernetes combined with container native storage allows enterprises to define their IT environments independent of infrastructure and cloud providers, and instantiate these software defined IT environment anywhere it is required.
The Data Challenge
The flexibility and portability afforded by this distributed version of the Enterprise cloud works as long as an application does not have a persistent state or data. In a stateful application, this flexibility disappears. Moving or copying a stateful application instance requires copying or migrating its associated data in the form of persistent volumes in Kubernetes. Traditional migration and data copy is time consuming, often requires significant downtime, is operationally complex and prevents customers from fully benefiting from the distributed pool of resources.
Data Teleport™ – Instant Data Mobility
Data Teleport mobilizes data and persistent volumes, making them portable as containerized applications. The functionality built into the Ionir storage and data management platform allows any persistent volume to be copied or moved between two Kubernetes clusters in less than 40 seconds, regardless of the size of the volume, the amount of data, or the location of the clusters. Data Teleport™ leverages Ionir’s unique metadata architecture to create a compact and secure portable representation of a volume and all its data that can be sent between two Kubernetes clusters. Once the portable representation is received by the target cluster, the volume and all of its data become immediately available. All write operations (which tend to be latency sensitive) are executed on the target, dismissing the need for invalidation on the source. For read operations, a combination of global deduplication and smart prefetch is used to minimize the read-on-demand and the overall amount of data transferred between the clusters.
Data Teleport™ can help make time consuming and often project-hybrid and multi-cloud workflows into everyday operations:
Production/Non-Production Hybrid. The current process of getting updated production data sets to either support development operation or debug production faults requires the entire data to be copied, which can take hours to days. With Data Teleport™, data sets can be copied and refreshed within 60 seconds, accelerating development and problem resolution.
Migrating Applications: Migrating or scaling applications requires significant downtime or staging of data, as well as advance planning. Data Teleport™ eliminates this complexity by allowing the application to start on a new Kubernetes cluster independent of whether it is in the same data center or on a different cloud with minimal downtime. The instant mobility operation copies “heat” information from the source for the volume to the target Kubernetes cluster. This heat information is used by the prefetch process to improve its predictions of what data will be required in the near future and hence needs to be prioritized.
Enterprises moving from traditional IT to Kubernetes-based clouds must be able to make their applications and data mobile if they want to maximize the flexibility and agility of their investments. Existing solutions that address the issue of data gravity based on data migration and data staging of persistent volumes are complex, time consuming, and in most cases require advance planning. Data Teleport™ is a fundamentally new capability that enables instant mobility of persistent volumes, creating a seamless distributed enterprise cloud in which applications can instantly and effortlessly move between infrastructure pools.