Ingres High Availability

Keep Mission-Critical Systems Online 24/7

High Availability means that systems and procedures ensure that mission critical systems remain on-line 24x7. This includes avoiding unplanned and planned downtime. High Availability ("HA") is closely related to Business Continuity ("BC").

Unplanned downtime includes power outages, failed CPU or RAM components or a fire in a data center. Ingres HVR protection from unplanned downtime typically consists of two elements;

• Business Continuity Server

  This is a local Ingres HVR replica of the production system, which gives fault tolerance if the main system is interrupted. It can also be used to avoid planned downtime.

• Disaster Recovery Server

  DR is a separate system which is located at another building, preferably in a different town.

Unplanned Downtime

• Offsite Disaster Recovery

  Because of Ingres HVR's extreme efficiency over a network it is affordable to install a disaster recovery database in another building or a different town. This protects from a power outage or network outage in the street, and also major disaster such as hurricane or earthquake.

Sometimes planned downtime can also carry the risk of unplanned downtime, for example after a major upgrade an important system may suddenly fail.

Load Balancing

• Active/Active

  The High Availability nodes connected together using Ingres HVR do not just have to be passive machines, which are only used after emergency. Instead they are active databases that can be exploited for spreading the load from the production machine. The transactions spread to the other machines can be read-only transactions or for a reporting database they can be read-write transactions for an active-active configuration.

Planned Downtime

Planned downtime is when the system must be brought offline because the scheduled activity cannot occur while end-users are connected. Traditionally planned downtime was scheduled at night, but this has become untenable as business has adapted to the internet and globalization. Planned downtime can be classified as one of the following;

• Maintenance

  Ingres HVR allows zero-downtime maintenance, which eliminates the need for a nightly or weekly batch window.

• Upgrades

  Ingres HVR allows zero-downtime upgrades of the Operating System, DBMS version and application.

• Migrations

  A migration means the reachitecturing of a system that can take a long period. Ingres HVR allows zero-downtime migrations to be done with minimized risk.