CA SSO configurations are pretty much the same irrespective of windows OR linux. But CA SSO configuration accounts for only 25% of the E2E Kerberos configuration. 75% of the configuration is outside CA SSO e.g. things like browser (IE vs Chrome vs FireFox vs cURL) settings for Kerberos, Hostname Resolution between servers, Keytab file version numbers, encryption handshake algorithms, Kerberos environment variable, Kerberos configuration files. Thus Kerberos is not about reading CA SSO documentation ONLY, but it is also about knowing what goes on outside SiteMinder.
Many a times we do get the comment that we have configured everything as per CA SSO documentation, but yet Kerberos is not working. Almost most of the time the issue is outside CA SSO Configuration (i.e. in the 75% configuration space e.g. things like browser (IE vs Chrome vs FireFox vs cURL) settings for Kerberos, Hostname Resolution between servers, Keytab file version numbers, encryption handshake algorithms, Kerberos environment variable, Kerberos configuration files).
what differs is the complexity outside siteminder? an e.g. In windows you can just added servers to the AD domains (which is also acting as KDC). In linux you have manually configure at OS level to resolve correctly & interwork with the AD domain.
Hope the thought process helps!