High Availability

The design and deployment of application infrastructure requires a High Availability model to ensure uptime and prevent impact from standard infrastructure maintenance. This can be accomplished by creating clusters of hypervisor nodes backed by redundant storage clusters. This is one of the many benefits seen by converting hardware to software in a virtualization model, as legacy dedicated infrastructure requires more specific configuration and application level redundancy to achieve the same result.

Altexxa Group utilizes numerous virtualization platforms for the design and deployment of High Availability systems. Enterprise class virtualization implementation also features automation within clusters to ensure applications shift between nodes to avoid failures.

  • Hypervisor clusters powered by redundant storage systems
  • Converting hardware to software by replacing legacy infrastructure
  • Automation systems within virtualization model for efficiency

Logical Diagram

  • Multiple virtualized node clusters for optimal failover availability
  • Cluster nodes directly connected to SAN layer switching
  • Redundant storage layer clustering with data replication

Infrastructure

High Availability virtual clusters should be deployed with high resource capacity to allow for multiple node failure without application impact. Infrastructure must also include standard redundancy considerations to eliminate single points of failure at the individual node layer. The High Availability model also includes the deployment of redundant storage clusters which provides protection from critical failure such as chassis or controller. Connectivity between the node layer and storage layer utilizes a redundant path SAN implementation. The logical implementation diagram will show a mirror infrastructure layout to ensure optimal availability.

  • Determining resource capacity requirements for failover protection
  • Node layer redundancy considerations to reduce impact risk
  • Clustered storage deployment - remove single point of failure
  • Mirror infrastructure layout with multiple network paths

Software Platforms

Altexxa Group works with hypervisor based virtualization solutions for High Availability infrastructure deployment. We primarily work with Citrix XenServer and VMware ESXi based virtualization and enabling the associated High Availability functionality. We have also been working with software orchestrator packages such as OpenStack within existing hypervisor clusters for High Availability management. Alert detection for High Availability events within node clusters is also handled with a Nagios monitoring deployment and is used to track failure patterns.

  • Enterprise class hypervisor software solutions for node clusters
  • Software orchestrator solutions for centralized HA management
  • Failure pattern tracking for proactive hardware investigation