I think I understand the token bucket scheme described in "Policing and Shaping Overview" as it applies to single rate policies. But when multiple committed access rate policies are used do they draw from a common token bucket or are they each given their own bucket?
Thank you. May I ask a followup, then. If each policer gets its own token bucket and extended bursting kicks in, are tokens borrowed from that policer's future allotment of tokens, or are they borrowed from other policers' unused tokens?
If tokens are not available the extended burst parameters and the compounded dept are examined to determine if the packet is transmitted or dropped, if allowed, tokens are then borrowed against that flows future allocation.
The following URL gives an in-depth explanation of Token buckets and how they are used, including how extended bursts are dealt with:
