https://tools.ietf.org/html/rfc3649
1. can it be simulated in packet tracer, i tried in 2911 but not support in this version
2. how to set ip tcp window-size to achieve lowest latency? what is this number?
3. if i set this window size, do i need to configure Microsoft window and Linux also using this window size?
4. if only router layer 3 need to configure, does it mean that only need to configure core switch in campus infrastructure then enough?
i use in RFC3649
| High Speed W = (p/Low_P)^S Low_Window |
| S = (log High_Window - log Low_Window)/(log High_P - log Low_P) |
5. but confused in (p/Low_P)^S Low_Window, i do not know whether it missed multiply operator
(p/Low_P)^S * Low_Window ?
6. is W = (p/Low_P)^S Low_Window , the window to be set in ip tcp window-size in router or window and linux?
7. how to calculate the RTT and relate with window size? as i know RTT is delay, the smallest the better?
A. it said packet drop at most 10^(-2) , then i set ((10^(-2))/Low_P)^S Low_Window , is it correct?
B. how to use excel to simulate the tradeoff and choose the best window size for lowest latency?
C. if core switch router can be do, why there is WCCP in ASA and WCCP Server behind ASA ? isn't this redirection makes larger latency?
D. what would happen if MS window and Linux not configured the same TCP window size as core switch router?
E. which certification in Cisco teach this calculation and about Cisco ACE and Cisco WCCP ?
