M86 Security SWG, User Manual, Page: 79 / 134

S W G U s e r G u i d e
79 Chapter 15: Implementing ICAP
4. To enable the ICAP client, select the Enable ICAP Client checkbox.
5. In the TimeOuts tab, adjust the following values, as needed:
• Connection Timeout — Maximum number of second to wait for a connection to be estab‐
lished. Default: 60.
• I/O Timeout — Maximum number of seconds to wait for completion of a message transmis‐
sion. Default: 120.
• Connection Reuse Timeout — Maximum number of seconds that the connection will be
alive on idle after its previous use. Default: 300.
6. In the Keep Alive tab, specify the number of seconds between each health check of ICAP Services.
The health check determines if the service is up and running. The default is 180 seconds.
7. Click Save .
8. If you are ready to distribute and implement the changes in your system devices, click .
Defining ICAP Service Groups
You must identify/define the ICAP Services that SWG as an ICAP client can request. However, each
ICAP Service must belong to an ICAP Service Group; therefore, before you can define an ICAP
Service, you must define the group to which it will belong.
Â
To define an ICAP Service Group
1. Select Policies
Æ
Condition Settings
Æ
ICAP Service Groups .
2. Do either of the following:
• To create an ICAP Service Group, right click the ICAP Service Groups (root) node, and
choose Add Group . The main window for defining the group is displayed.
• To edit an existing group, select the group node, and in the main window click Edit .
The ICAP Service Group window is displayed. This window contains two tabs:
•
General tab — for defining basic parameters of the group.
• Advanced tab — for defining timeout parameters for the group.
3. Fill in the General tab, as follows:
a. In the Name field, specify a name for the ICAP Service Group.
b. In the Method field, select the mode in which the ICAP protocol works. Note : Currently, only
one mode is supported.
•
REQMOD — Request Modification. This mode processes the Client request to a distant
server while it is being sent to the internet, that is, before it reaches the internet. An
example of Request Modification is a Data Leakage Prevention (DLP) server that scans all
outgoing requests and enforces the organizations DLP policy.
c. In the Load Balancing Algorithm field, select the type of algorithm that should be used for
distributing the load. Note : Currently, only one algorithm is supported.
• Round Robin — Distributes the load between servers sequentially in a circular pattern.