Repeat the following
procedure for each appliance in the cluster:
- Fill in the WCCP
- Navigate to
Enable to enable WCCP mode on the appliance.
(Multiple Caches) option.
- Fill in parameters in the
CloudBridge Cluster section.
- Enter T5 from your worksheet as the Cache 1 IP, T6 as the Cache
2 IP, T2 as the Subnet Mask, and T1 as the Gateway. Click
Service Group section appears.
- In the Service Group
Details section, specify the WAN and LAN service groups (T11 and T12 from your
- In the Priority field,
practice this value is somewhat arbitrary).
- From the Protocol list,
- In the DC Algorithm field,
Disruptive. “Deterministic” is always safe to use, and should be used if
you are using only two appliances, or are using multiple routers. “Least
Disruptive” disrupts fewer user sessions on failover when used with clusters of
three or more appliances, but has restrictions on its use.
Service Group Pair Status to On.
- If your router is
configured to require a password, enter the password in the
Password field. Otherwise, leave the field blank.
- In the
Router Communications Details section, enter the IP address of
the router (T8 on your worksheet: often identical to T1 as well). This is the
IP address of the appliance-facing router interface. If you use multiple
routers to communicate with the appliance, list them all here.
- From the
Router Forwarding list, select Level 2 or GRE, according to the capabilities of
your router. Use Level 2 if you can, and GRE if you must.
- For the Mask Value, enter
the value you determined from the WCCP Clustering worksheet. This is a critical
value: a poor choice will result in poor load-balancing or none at all.
Create. This creates the WAN and LAN service groups.
- On the
page, change the bandwidth limits on each
defined WAN to 95% of the aggregate speed of all your WANs. This prevents the
link from being under-utilized when load-balancing is imperfect. If ICA
(XenApp/XenDesktop) is the dominant use, set each appliance to (95% of WAN
bandwidth)/N, where N is the number of appliances (or twice the number of
appliances if they are CloudBridge 4000 or 5000 units), to divide the bandwidth
equally among the appliances. This latter method is most appropriate for
applications with large numbers of active connections that have relatively low