High availability and disaster recovery

From Genesys Documentation
Jump to: navigation, search

High availability (HA) and disaster recovery (DR) are two important factors in establishing a resilient infrastructure. This article describes the two supported architecture types for HA and DR, as well as the HA and DR modes supported by the private edition services.

Early Adopter Program
Genesys Engage cloud private edition is being released to pre-approved customers as part of the Early Adopter Program. Please note that the documentation and the product are subject to change. For more details about the program, please contact your Genesys representative.

Modern software environments demand two major types of agility:

  • The ability to autoscale—that is, to rapidly increase processing power to handle a growth in interaction volume
  • Resiliency—that is, the ability to fail over after losing one or more services—or even a whole data center or region

The second type of agility—the ability to bounce back from a failure—is broadly divided into two types of activity, each with its own requirements:

  • High availability (HA) is the use of built-in redundancy to handle the failure of a service within a single region or data center
  • Disaster recovery (DR) is the ability to continue processing after losing a whole region or data center, by failing over to another region or data center

Genesys Engage cloud private edition allows you to set up a highly available and resilient infrastructure whether you are using a cloud deployment or hosting it in a private data center.

Note, however, that these two types of deployments require somewhat different architectures, as discussed below.

Important
Before you continue, review the platform section of the private edition architecture page for an in-depth discussion of key components of the private edition architecture, such as unit pairs and Availability Zones.

Key architectural distinctions

Both the cloud and private data center architectures use multiple geographical regions that are hosted within a single unit group. And in both types of environment, all of the unit pairs in a deployment are fully meshed with each other.

But the cloud deployment's ability to use Availability Zones makes its redundancy features more robust, as shown in the following table:

Deployment type Redundancy type Characteristics
Cloud Availability Zones within regions Multiple data centers in a small geographical area—can share a single Kubernetes cluster
Private data center Physically discrete data centers Data centers cannot share a Kubernetes cluster


Cloud architecture

One of the most important advantages of a cloud architecture is the enhanced redundancy through the use of Availability Zones (AZs). As discussed in the platform section of the private edition architecture page, an AZ is a discrete location within a region that is designed to operate independently from the other Availability Zones in that region. Because of this separation, any given Availability Zone is unlikely to be affected by failures in other Availability Zones.

In the cloud architecture, high availability is achieved by deploying instances within different Availability Zones.

Important
Black pod icons indicate services that can only be hosted in the primary unit.
Pe-units-group-internals-cloud.png

Private data center architecture

Important
Black pod icons indicate services that can only be hosted in the primary unit.
Pe-units-group-internals-private-dc.png

Planning for high availability

Private edition services scale automatically to meet demand. And when a service fails, private edition's high availability features enable an auto restart of that service.

For first-time deployments, you must plan:

  • The number of nodes
  • The number of pods that each node must run in your Kubernetes cluster

In order to reduce service disruptions, Genesys recommends that you run a minimum of three pod replicas for each service. Use the Sizing Calculator to determine the infrastructure requirements for achieving high availability in your contact center.


Resiliency modes of private edition services

High availability modes

Private edition services maintain high availability by using the following modes:
Important
Some services support more than one HA mode.
High availability modes
Mode Description
N = 2 (active-active) The service is running on two nodes simultaneously. If one fails, the other takes over.
N = 1 (singleton) The service is running on a single node. If that node fails, a new node is started to take over processing for that service.
N = N (N+1) The service normally runs on N nodes. If a node fails, a new node is started to replace the failing node.
Cron jobs Some services run as cron jobs, meaning that normal HA is not applicable

Disaster recovery modes

Private edition services achieve disaster recovery by using the following modes:

Disaster recovery modes
Mode Description
Active spare A complete production replica is in place and serves traffic during normal operations
Limited active spare A complete production replica is in place and serves traffic during normal operations, but the data is only used in case of disaster
Pilot light The bare minimum configuration is in place to get the system back within a short time period. For example, there might be a read replica for a database. Application servers and web servers are deployed after the disaster.
Not supported Disaster recovery is not supported for this service

Modes for each service

The following table displays the high availability and disaster recovery modes used by private edition services.
Important
Disaster recovery is not supported for services that are only available in the primary unit.

This information is under development: Flagged items aren't yet confirmed or have info coming soon; Checked items are valid.

Service & Included ServicesHigh AvailabilityDisaster RecoveryWhere can you host this service?

Genesys Authentication

N = N (N+1)

Billing Data Service

BDS is a Cronjob that runs on a per-tenant basis, so High Availability (HA) is not applicable.

Genesys Voice Platform

  — Voice Platform Configuration Server N = 1 (singleton) active-active
  — Voice Platform Media Control Platform N = N (N+1) active-active
  — Voice Platform Reporting Server N = 1 (singleton) active-active
  — Voice Platform Resource Manager N = 2 (active-active) active-active
  — Voice Platform Service Discovery N = 1 (singleton) active-active

Workspace Web Edition

N = N (N+1)

Active-spare

Genesys Engagement Service

N = N (N+1)

Genesys Customer Experience Insights

N = 2 (active-active)

active-active

Genesys Info Mart

N = 1 (singleton)

active-active

Tenant Service

N = N (N+1)

active-active

Telemetry Service

N = N (N+1)

Universal Contact Service

N = N (N+1)

active-backup

Voice Microservices

N = N (N+1)

active-active