Method of transmitting data
First Claim
Patent Images
1. A method of transmitting data from customers over a computer network in packets, the method comprising:
- marking each of said packets by a first state or a second state by a network device;
wherein said first state has a high drop precedence and is based on a random probability (p) calculated with reference to a previous random probability and a previous token bucket occupancy;
wherein the packet is marked by a first state if a length of the packet is above a token bucket occupancy of a token bucket; and
wherein the probability (p) at a given step is expressed as
p=k1×
(bref−
b)−
k2×
(bref−
bold)+pold and that at a next step pold is set equal to p and bold is set equal to b, wherein pold and bold are values that are respectively the values p and b had at a previous update time,bref is a desired token bucket occupancy,b is a total bucket size,k1 is a constant, andk2 is a constant.
1 Assignment
0 Petitions
Accused Products
Abstract
A method of transmitting data from customers (C1, C2, C3, C4, C5, C6, C7, C8, C9, C10) over a computer network, in particular over the Internet, where the data to be sent is split into packets, in particular into IP packets, where each packet is marked by one of at least two states (IN, OUT) and where the states (IN, OUT) determine which packets are dropped first, if packets are dropped during transmission, is, with regard to optimizing the drop rate of the packets, characterized in that the marking of the packet with a state of high drop precedence (OUT) is based on a random probability (p).
-
Citations
42 Claims
-
1. A method of transmitting data from customers over a computer network in packets, the method comprising:
-
marking each of said packets by a first state or a second state by a network device; wherein said first state has a high drop precedence and is based on a random probability (p) calculated with reference to a previous random probability and a previous token bucket occupancy; wherein the packet is marked by a first state if a length of the packet is above a token bucket occupancy of a token bucket; and wherein the probability (p) at a given step is expressed as
p=k1×
(bref−
b)−
k2×
(bref−
bold)+poldand that at a next step pold is set equal to p and bold is set equal to b, wherein pold and bold are values that are respectively the values p and b had at a previous update time, bref is a desired token bucket occupancy, b is a total bucket size, k1 is a constant, and k2 is a constant. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42)
wherein R(t) is a round trip time as a function of time, and W(t) is a TCP window size as function of time.
-
-
20. The method of any of claims 1 or 2, further comprising stabilizing the token bucket occupancy by linearizing the change in a token bucket occupancy, wherein the change in the token bucket occupancy (b) is expressed as
-
( t ) R ( t ) N ( t ) + C , wherein N(t) is a number of TCP sources of the customer as a function of time, W(t) is a TCP window size as function of time, R(t) is the round trip time as a function of time, and C is an assigned maximum bandwidth.
-
-
21. The method of claim 19, wherein the change in a TCP window size (W) is linearized in an operation point, preferably at constant round trip time (Ro) or constant number of TCP sources (N) expressed as
-
W . = N R 0 2 C ( δ W + δ W ( t - R 0 ) ) - R 0 C 2 2 N 2 δ p ( t - R 0 ) , δ b . = - N R 0 δ W , where
δ
w=w−
wo
δ
b=b−
bo
δ
p=p−
po,C is an assigned maximum bandwidth, b is a token bucket occupancy, W is a TCP window size, and wo, bo, and po are determined by setting W=0 and b=0.
-
-
22. The method of any of claim 1 or 2, wherein the token bucket occupancy (b) is stabilized by way of a controller whose transfer function is expressed as
-
( s ) = K s z + 1 s , wherein C(s) is a controller transfer function, s is a LaPlace parameter, z is a zero of the controller, and K is a constant.
-
-
23. The method of claim 22,
wherein a zero of the controller is to -
g = 0.1 2 N - R + 2 C in order to have the controller dominate the closed-loop behavior, wherein ω
g is a maximum frequency,N−
is a minimum in a range for a number of TCP sources where N≧
N−
,R+ is a maximum range of round-trip time (Ro) where Ro≦
R+, andC is an assigned maximum bandwidth.
-
-
24. The method of claim 22, wherein especially by invoking a Nyquist criterion a gain (K) in the controller is set to
-
( 2 N - ) 3 ( 2 R + C 2 ) 2 , wherein N−
is a minimum in a range for a number of TCP sources where N≧
N−
,R+ is a maximum range of round-trip time (Ro) where Ro≦
R+, andC is an assigned maximum bandwidth.
-
-
25. The method of claim 1, wherein k1 is computed by a bilinear transformation expressed as
-
( T 2 + 1 ω g ) , wherein K is a constant, T is a sampling time, and wg is an inverse of a control constant.
-
-
26. The method of claim 1, wherein a k2 is computed by a bilinear transformation expressed as
-
( T 2 - 1 ω g ) , wherein K is a constant, T is a sampling time, and wg is an inverse of a control constant.
-
-
27. The method of claim 1, wherein the computers of the network are linked with each other.
-
28. The method of claim 5, wherein the links have maximum bandwidths or the customers have been provided an assigned maximum bandwidth for the purpose of data transmission.
-
29. The method of claim 7, wherein the links have maximum bandwidths or the customers have been provided the assigned maximum bandwidth for the purpose of data transmission.
-
30. The method of claim 7, wherein the marking of the packet is based on the comparison of the current bandwidth with the assigned maximum bandwidth.
-
31. The method of claim 8, wherein the marking of the packet is based on the comparison of the current bandwidth with the assigned maximum bandwidth.
-
32. The method of claim 7, wherein the packet is marked with a high drop precedence if the current bandwidth is higher than the assigned maximum bandwidth.
-
33. The method of claim 8, wherein the packet is marked with a high drop precedence if the current bandwidth is higher than the assigned maximum bandwidth.
-
34. The method of claim 9, wherein the packet is marked with a high drop precedence if the current bandwidth is higher than the assigned maximum bandwidth.
-
35. The method of claim 10, wherein the packet is marked with a high drop precedence if the current bandwidth is higher than the assigned maximum bandwidth.
-
36. The method of claim 20, wherein the change in the token bucket occupancy (b) is linearized in an operation point, preferably at constant round trip time (Ro) or constant number of TCP sources (N) expressed as
-
W . = N R 0 2 C ( δ W + δ W ( t - R 0 ) ) - R 0 C 2 2 N 2 δ p ( t - R 0 ) , δ b = - N R 0 δ W , where
δ
w=w−
wo
δ
b=b−
bo
δ
p=p−
po″
, andC is an assigned maximum bandwidth, W is a TCP window size and wo, bo, and po are determined by setting W=0 and b=0.
-
-
37. The method of claim 23, wherein by invoking a Nyquist criterion a gain (K) in the controller is set to
-
( 2 N - ) 3 ( 2 R + C 2 ) 2 • , wherein N−
is a minimum in a range for a number of TCP sources where N≧
N−
,R+ is a maximum range of round-trip time (Ro) where Ro≦
R+, andC is an assigned maximum bandwidth.
-
-
38. The method of claim 23, wherein a k1 is computed by a transformation expressed as
-
1 = K ( T 2 + 1 ω g ) • , wherein K is a constant, T is a sampling time, and wg is an inverse of a control constant.
-
-
39. The method of claim 24, wherein k1 is computed by a transformation expressed as
-
1 = K ( T 2 + 1 ω g ) • , wherein K is a constant, T is a sampling time, and wg is an inverse of a control constant.
-
-
40. The method of claim 23, wherein k2 is computed by a transformation expressed as
-
2 = - K ( T 2 - 1 ω g ) • wherein K is a constant, T is a sampling time, and wg is an inverse of a control constant.
-
-
41. The method of claim 24, wherein k2 is computed by a transformation expressed as
-
2 = - K ( T 2 - 1 ω g ) • , wherein K is a constant, T is a sampling time, and wg is an inverse of a control constant.
-
-
42. The method of claim 25, wherein k2 is computed by a preferably bilinear transformation expressed as
-
2 = - K ( T 2 - 1 ω g ) • , wherein K is a constant, T is a sampling time, and wg is an inverse of a control constant.
-
Specification