Fail-over Web Services

Introduction

Failover the Web Services is only valid for configurations with multiple A365 servers. The failover can take place within the same Front End pool, where two A365 servers are part of an enterprise trusted application pool.

Figure 1 Single Front-End pool with a failover server

Another supported configuration is when there are two paired Front-End pools, where a single A365 server is connected to each pool.

Figure 2 Two paired Front-End pools with a failover server

The most advanced configuration is where you have two paired Front-End pools, each with an enterprise trusted application pool.

Figure 3 Two paired Front-End pools each pool with a failover server

Anywhere365 client applications like Attendant, Snapper, Wallboard, Web Wallboard and Webchat communicate with the Web services of Anywhere365. The Web services communicates with the UCC A Unified Contact Center, or UCC, is a queue of interactions (voice, email, IM, etc.) that are handled by Agents. Each UCC has its own settings, IVR menus and Agents. Agents can belong to one or several UCCs and can have multiple skills (competencies). A UCC can be visualized as a contact center “micro service”. Customers can utilize one UCC (e.g. a global helpdesk), a few UCC’s (e.g. for each department or regional office) or hundreds of UCC’s (e.g. for each bed at a hospital). They are interconnected and can all be managed from one central location. service on the same server.

If the UCC service is no longer running on a server, but the Web services still are, then they no longer provide information to the clients.

To make the failover faster and easier. It is best to use dedicated hostname (FQDN FDQN stands for Fully Qualified Domain Name and specifies the exact location in the tree hierarchy of the Domain Name System (DNS). An example for [hostname].[domain].[top level domain] is [www].[microsoft].[com]) for the UCC host and create a DNS record for it and configure the clients to use the dedicated hostname. This can be a CNAME record pointing to the active A365 server name (or an A record pointing to the server IP address). The DNS record can be resolved by internal clients.

Scenarios

Below are two possible scenarios to make switching webservice access from one active server to another server easier.

DNS record without a load balancer

There is no automatic failover for the web services in this scenario. An Administrator needs to modify the DNS record used for the dedicated hostname and clients need to resolve the DNS change as soon as possible.

Configuration

Determine the dedicated hostname for the Anywhere365 web services. As a sample: a365web.<domain>.

Create a DNS record of type CNAME for the dedicated hostname and point it to the FQDN of the active A365 server. Set the time to live (TTL) as short as possible. Use the dedicated hostname as the UCC host for the configuration of the clients.

Sample in the Snapper:

Failover to passive server

In case of a failover modify the DNS record and change the FQDN to the FQDN of the passive server.

Failback to active server

In case of a failback modify the DNS record and change FQDN to the FQDN of the primary server.

DNS record with a load balancer

The advantage of using a load balancer is that the failover of the web services in most case takes place automatically. A manual failover might be necessary in case of a datacenter failover with load balars in both datacenters.

Configuration

Determine the dedicated hostname for the Anywhere365 web services. As a sample: a365web.<domain>.

Determine the virtual IP address for the load balancer. Create a DNS record for the dedicated hostname and point it to the virtual IP address or use a CNAME record in case of global load balancing (two datacenters).

Create a virtual service for the Anywhere365 web services. Add the active and passive Anywhere365 servers as real servers to the virtual service. Determine which UCC is used for the health check and use the port the trusted application for the health check. Make sure that this UCC will not be removed in the future, if so then change the health check port to another trusted application. Configure the health check: TCP port (sample: 6001).

If the health check rule returns OK, then UCC service is running on the server. If it returns NOK, then the UCC service is not running on the server.

Use the dedicated hostname as the UCC host for the configuration of the clients.

See sample in the Snapper:

Failover to passive server

No action is required. Clients connect to the virtual service on the load balancer and the load balancer connects the clients to the passive server where the UCC service is running. In case of global load balancing across two datacenters, the CNAME record needs to be changed, so that it points to the active datacenter.

Failback to active server

No action is required. Clients connect to the virtual service on the load balancer and the load balancer connects the clients to the active server. In case of global load balancing across two datacenters, the CNAME record needs to be changed, so that it points to the active datacenter.

Multiple trusted application pools

As mentioned briefly earlier, when you have multiple enterprise trusted application pools, then you also need multiple dedicated hostnames and multiple virtual services (if using a load balancer or multiple load balancers).

Sample:

Front End pool	Trusted application pool	Active server	Passive server	Dedicated hostname
fepool1.<domain>	a365pool1.<domain>	SRV1	SRV2	a365web1.<domain>
fepool2.<domain>	a365pool2.<domain>	SRV3	SRV4	a365web2.<domain>

UCC on trusted application pool 1

UCC on trusted application pool 2

Clients

Web Wallboard

During the installation of the Web Wallboard, you need to provide the URLs for the Attendant and the Dashboard service. Instead of using the FQDN of the application server, use the FQDN of the dedicated hostname.

Snapper

During the installation of the Snapper, you to provide a UCC name and the UCC host name. Instead of using the FQDN of the application server for the UCC name, use the FQDN of the dedicated hostname.

Attendant

During the installation of the Attendant, you to provide a UCC name and the UCC host name. Instead of using the FQDN of the application server for the UCC name, use the FQDN of the dedicated hostname.

Inflight Wallboard

During the installation of the Inflight Wallboard, you to provide a UCC name and the UCC host name. Instead of using the FQDN of the application server for the UCC name, use the FQDN of the dedicated hostname.

Webchat Plus

During the installation of Webchat Plus, you are asked to provide the FQDN of the server. Instead of using the FQDN of the application server, use the FQDN of the dedicated hostname.

SSL

If you are using HTTPS then make sure that the FQDN belonging to the dedicated hostname is also included in the certificate on the A365 servers.