About TAC and High Availability

High availability is most important for any organizations to keep the business operation up and prevent service down by any means.

There are many ways to implement high availability.

TAC support cloud-cloud, cloud-on-premises and on-premises - on-premises setups.

TAC does not have any subnet limitations or availability zone limitations that are a requirement for cloud platforms. Licensing does not based on cloud infrastructure and it does not going to affect licensing. Customer needs to decide how they are going to deploy TAC in their network.

Following are the high availability deployment methods:

  1. Join TAC instances into an array –
    1. This is the recommended approach by PortSys.
    2. TAC array members utilize license from TAC array master and doesn't require dedicated license for each instances.
    3. Array master and Array members share same TAC configuration.
    4. There should be proper network connectivity between on-premises- on-premises or cloud and on-premises network or cloud-cloud network to let TAC communicate with each other depend on the setup.
    5. On multi-NIC setup it is recommended to add Internal NIC first and make it private network (or domain joined) before adding the External NIC. This is to overcome Microsoft's Network Location Aware (NLA) to properly pick appliances NIC to proper NLA category.
    6. In any type of HA setup all TAC array members must be joined to the same Windows domain.
    7. There shouldn't be any Firewall blockage in between TAC array nodes on the private network.

  2. TAC in Array + DR ( 2 + 1 ) - 
    1. Same as point # 1. 
    2. DR TAC server/s will not participate in integrated NLB operation with main TAC array nodes.
      Read more  about deploying TAC in Array  + DR .
    3. DR TAC Uses Dedicated IP to connect from outside and the DR fail over is a manual operation.
    4. There shouldn't be firewalls in between array master and DR servers for array communication in private network. Having Firewall may lead to unexpected array communication issues because of dynamic port usage in array sync.

  3. TAC in Geo load balancing (individual servers with Geo LB) –
    1. require separate licensing for each instances.
    2. Instances won't share same configuration and need to configure manually.
    3. They can place both instances in same subnet of cloud or one in cloud and another instance in on-site.
    4. This can be managed by placing a geo load balancer (DNS load balancer) in front of TAC to handle incoming web requests.
  4. TAC in Disaster recovery (individual servers) –
    1. same as point #2. 2nd instances uses only as DR server.
    2. require separate licensing for each instances.
    3. Instances won't share same configuration and need to configure manually.
    4. They can place both instances in same subnet of AWS cloud or one in AWS and another instance in on-site.
    5. This can be managed by placing a geo load balancer in front of TAC to handle incoming web requests.