US 10,771,319 B1 · Robustness verification method and apparatus for distributed control plane in software-defined network

Abstract

Provided are a robustness verification method and apparatus for a distributed control plane in a software-defined network, including constructing a robustness verification framework, where the robustness verification framework accommodates a failure scenario set and a failure recovery strategy set; selecting a failure recovery strategy, and querying a worst failure scenario under the failure recovery strategy; verifying a utilization rate of a biggest controller in the control plane under the combination of the worst failure scenario and the failure recovery strategy; and verifying the robustness of the control plane based on the utilization rate. The verification problem of the control plane is taken as a robustness optimization problem under different failure cases and failure recovery strategies. After the failure recovery strategy is selected, the worst failure scenario is obtained to determine whether the performance of the control plane satisfies requirements.

Claims12

Dependent claims

A robustness verification method of a distributed control plane in a software-defined network, comprising:

at block S1, constructing a robustness verification framework, wherein the robustness verification framework accommodates a failure scenario set and a failure recovery strategy set;
at block S2, selecting a failure recovery strategy and querying a worst failure scenario under the failure recovery strategy;
at block S3, verifying a utilization rate of a biggest controller in the control plane under the combination of the worst failure scenario and the failure recovery strategy;
and
at block S4, performing verification for robustness of the control plane based on the utilization rate;
wherein the formula used for performing verification for the robustness of the control plane at block S4 is as follows:
max
⁢
w
^
⁡
(
v
)
s
.
t
.
⁢
{
∑
j
=
1
M
⁢
⁢
v
j
f
≤
f
∑
i
=
1
N
⁢
⁢
∑
j
=
1
M
⁢
⁢
v
ij
l
≤
l
{
v
j
f
,
v
ij
l
}
∈
{
0
,
1
}
wherein N is a number of switches in the control plane, M is a number of controllers in the control plane, vlij refers to a state of a secure link between the switch si and the controller cj, vfj refers to a state of the controller cj, and max ŵ(v) is a failure recovery strategy.

64 more elements

A robustness verification method of a distributed control plane in a software-defined network, comprising:

at block S1, constructing a robustness verification framework, wherein the robustness verification framework accommodates a failure scenario set and a failure recovery strategy set;
at block S2, selecting a failure recovery strategy and querying a worst failure scenario under the failure recovery strategy;
at block S3, verifying a utilization rate of a biggest controller in the control plane under the combination of the worst failure scenario and the failure recovery strategy;
and
at block S4, performing verification for robustness of the control plane based on the utilization rate;
wherein the manner of verifying the robustness of the control plane based on the utilization rate at block S4 is determining a relationship of F* and 1, and if F*<1, it is considered that the control plane satisfies design requirements and otherwise capacity expansion is performed for the control plane,
wherein the capacity expansion is performed based on the following capacity expansion criterion:
min
δ
≥
0
⁢
max
v
∈
V
⁢
min
⁢
{
∑
j
=
1
M
⁢
⁢
φ
j
⁢
δ
j
❘
U
⁢
⁢
ρ
j
⁡
(
1
-
v
j
f
)
≥
∑
i
=
1

S
-
S
f

⁢
⁢
r
i
⁢
m
ij
*
+
∑
x
=
1

S
f

⁢
⁢
r
x
⁢
m
xj
b
xj
⁡
(
1
-
v
xj
l
)
≥
m
xj
∑
j
=
1
M
⁢
⁢
m
xj
=
1
}
,
and
wherein V is a failure scenario set, v is a failure scenario, δj is the biggest utilization rate of the controller, φj is a unit capacity cost of the controller, vlij refers to a state of secure link between the switch si and the controller cj, vfj refers to a state of the controller cj, ρj refers to a cost for increasing one unit of capacity for the controller cj, S refers to a set of switches, Sf refers to a set of failed switches, |Sf| refers to a number of the failed switches, mxj refers to whether the controller cj is the master controller of the switch si, and M refers to a number of controllers of the control plane.

106 more elements

A robustness verification system for a distributed control plane in a software-defined network, comprising:

a constructing module, configured to construct a robustness verification framework, wherein the robustness verification framework accommodates a failure scenario set and a failure recovery strategy set;
a selecting module, configured to select a failure recovery strategy and query a worst failure scenario under the failure recovery strategy;
an obtaining module, configured to verify a utilization rate of a biggest controller in the control plane under the combination of the worst failure scenario and the failure recovery strategy;
and
a verifying module, configured to verify robustness of the control plane based on the utilization rate;
wherein the formula used by the verifying module to verify the robustness of the control plane is as follows:
max
⁢
w
^
⁡
(
v
)
s
.
t
.
⁢
{
∑
j
=
1
M
⁢
⁢
v
j
f
≤
f
∑
i
=
1
N
⁢
⁢
∑
j
=
1
M
⁢
⁢
v
ij
l
≤
l
{
v
j
f
,
v
ij
l
}
∈
{
0
,
1
}
wherein N is a number of switches in the control plane, M is a number of controllers in the control plane, vlij refers to a state of a secure link between the switch si and the controller cj, vfj refers to a state of the controller cj, and max ŵ(v) is a failure recovery strategy.

64 more elements

Dependencies are inferred from the claim text.

Robustness verification method and apparatus for distributed control plane in software-defined network

Life of this patent

Robustness verification method and apparatus for distributed control plane in software-defined network

Life of this patent

Abstract

Claims12