Unlock Kubernetes and Simplify Operations: Introducing YugabyteDB 2.18
Meet YugabyteDB 2.18, designed to accelerate your data modernization efforts with new cloud native features, like powering multi-region Kubernetes at scale and supporting a diverse set of workloads. You can also simplify operations with new automation and intelligence added to our powerful self-managed DBaaS solution, YugabyteDB Anywhere.
Continue reading for details of these new features, including colocated tables, multi-region Kubernetes, YugabyteDB Anywhere Performance Advisor, new query pushdowns to further enhance PostgreSQL performance, and more!
Organizations need to deploy global applications that require a global app infrastructure and a geo-distributed data layer. Kubernetes has become the default choice for many modern applications.
Kubernetes enhancements in YugabyteDB 2.18 make it easier for enterprises to deliver Kubernetes with distributed SQL databases at scale. A range of new features now available in YugabyteDB Anywhere help with seamless multi-region deployments, make the initial configuration easier, provide faster and more granular data protection, and support new ecosystem platforms.
YugabyteDB 2.18 delivers five key enhancements in YugabyteDB Anywhere for Kubernetes environments.
- Multi-region Kubernetes with Synchronous Replication: Enhance application availability and resiliency of applications on YugabyteDB and Kubernetes with the addition of multi-region, stretched cluster support using the Kubernetes Multi-Cluster Services (MCS) API. This is easily configured through the YugabyteDB Anywhere interface, so you can take advantage of a powerful deployment architecture. The feature is currently available as an Early Access offering and supported for production deployments.
- xCluster Asynchronous Replication on Kubernetes: YugabyteDB Anywhere now supports the deployment and management of xCluster asynchronous replication on Kubernetes. Combined with our previous update, you now have even more deployment choices that are easy to configure and manage. Take advantage of YugabyteDB’s unique xCluster capabilities for two data-center deployments, to support, for example, disaster recovery strategies.
- Simplified Kubernetes Deployment: Kubernetes can introduce complexity to your environment, but YugabyteDB 2.18 delivers a number of updates that simplify the process of deploying and aligning the Yugabyte database with your preferred Kubernetes configuration. Improvements include non-root support, shared namespaces across universes, and the ability for the database and Kubernetes environment to share a cluster and namespace.
- 5x Faster Backup and Restore Processes: Your Kubernetes environments can now take advantage of the backup mechanism introduced in YugabyteDB 2.16—the YugabyteDB Controller, which runs as a container on each Kubernetes pod. The new engine accelerates backup and restore processes up to 5x faster than before. As your cluster scales, the backup capabilities scale linearly. In addition, file-level incremental backups are now supported with Kubernetes, helping create backups up to 20x more frequently.
- Ecosystem Expansion with Mirantis: Mirantis Kubernetes Engine (MKE) is a container orchestration platform for developing and running modern applications at scale on private clouds, public clouds, and bare metal. MKE is often used when orchestrating more than one container or when using multi-tenant software offerings. YugabyteDB cluster can now be confidently deployed on MKE using kubectl and helm.
Today organizations of all sizes must support a growing set of diverse applications. These apps are constantly changing due to unpredictable customer and macroeconomic trends. How do you ensure you can support any app, while still delivering strong performance and simplified operations?
Most major cloud providers take a “portfolio approach,” meaning they offer a range of databases, each designed for different workloads, that you choose between, migrate between, and manage separately. YugabyteDB 2.18 continues to deliver key features that allow us to support diverse workloads on a single relational database.
New features for diverse workload support include:
- Colocated Tables: Colocated tables allow users to store (or co-locate) data from numerous small, related tables onto a single YugabyteDB tablet (i.e. shard), boosting performance and eliminating query fan-out. If an application or table grows, YugabyteDB intelligently adjusts without requiring complex migration, manual steps, or database silos. This provides seamless and intelligent support for a wide range of applications. Colocation is intended for small datasets and can boost performance while maintaining an automatic scale-out architecture across multiple shards for larger tables where needed. Based on your workloads, or how they change, you have the power to colocate a table on a per-table basis.
- Query Pushdowns: YugabyteDB continues to evolve beyond the core capabilities of PostgreSQL by taking advantage of YugabyteDB’s custom-built, distributed storage layer. YugabyteDB optimizes the performance of SQL queries by moving the processing into the distributed storage layer—as close to the data as possible. YugabyteDB 2.18 supports distinct pushdowns for index columns, pushdowns for the NEXTVAL function for sequences, and enables expression pushdowns by default.
- Scheduled Full Compactions: Background data compactions are critical to maintaining an efficient, high performance database. However, there are several scenarios that require more aggressive action than the normal automated process. Full compactions are needed for many scenarios, including those with large numbers of updates/deletes, workloads with TTL (time-to-live) enabled, or if compliance requires full deletion at a certain time. Historically, this had to be manually configured, but with YugabyteDB 2.18 you can schedule full compactions (with an adjustable time period that defaults to 30 days).
Many organizations are looking toDBaaS for its built-in automation and intelligence. However, some organizations have requirements, such as corporate security policies, that limit their cloud adoption ability. These organizations may want (or need) to retain control of the underlying infrastructure, security, and/or ongoing operations.
Even if they can’t adopt the cloud, these teams, who are often limited in terms of resources, need to embrace built-in intelligence and automation to efficiently support a growing set of initiatives. YugabyteDB Anywhere provides a powerful self-managed DBaaS solution for organizations that are not ready to fully embrace a cloud-first strategy.
With this release, YugabyteDB Anywhere delivers greater levels of automation, intelligence, and protection, including:
- Performance Advisor: YugabyteDB Anywhere users now have access to a powerful Performance Advisor that can scan your clusters to discover key performance optimizations. This built-in tool provides guidance on schema design, connections, indexes, query load, and CPU skew and usage. The advisor helps diagnose query performance issues and identify nodes that have become hot spots.
- Terraform Provider: Operators can now automate the deployment of YugabyteDB Anywhere clusters through the new Terraform Provider, helping them consistently deploy and manage clusters, and reduce time-to-value. Available as an Early Access offering, the Terraform Provider for YugabyteDB Anywhere is currently supported for production deployments.
- Advanced Backups and Protection: Confidently protect business-critical data through 5x faster backup and restores, point-in-time recovery (PITR), and new incremental backups—all now generally available in YugabyteDB 2.18. Incremental backups reduce the amount of data that needs to be transferred per backup by only looking at the delta, enabling more frequent backups at a lower cost.
YugabyteDB 2.18 is available today! All core YugabyteDB features are available for free in the 100% open source version, and YugabyteDB Anywhere orchestration engine enhancements are available to commercial customers.
Visit the following resources for more information: