记录下关于论文的理解，论文读的一知半解的。
在段路由中，段和段之间依然遵循最短路径的原则。在不存在段路由的情况下，
f
o
r
w
(
x
→
y
)
forw(x\rightarrow y)
forw(x→y)=SP(x,y)。在段路由存在的情况下，
f
o
r
w
(
x
1
→
.
.
.
→
x
k
)
=
∪
i
=
1
k
−
1
f
o
r
w
(
x
i
→
x
i
+
1
)
forw(x_1\rightarrow ...\rightarrow x_k)=\cup_{i=1}^{k-1}forw(x_i\rightarrow x_{i+1})
forw(x1→...→xk)=∪i=1k−1forw(xi→xi+1)。对于一条路径
p
⃗
=
f
o
r
w
(
x
1
→
.
.
.
→
x
k
)
，
\vec p=forw(x_1\rightarrow ...\rightarrow x_k)，
p=forw(x1→...→xk)，在某些边因为故障失效后，段和段之间会重路由。
f
f
f是一条失效的边，
f
∈
F
f\in \mathscr{F}
f∈F。
f
o
r
w
(
p
⃗
,
f
)
=
∪
i
=
1
k
−
1
f
o
r
w
(
x
i
→
x
i
+
1
,
f
)
=
∪
i
=
1
k
−
1
S
P
(
x
i
,
x
i
+
1
,
f
)
forw(\vec p, f)=\cup_{i=1}^{k-1}forw(x_i\rightarrow x_{i+1},f)=\cup_{i=1}^{k-1} SP(x_i,x_{i+1},f)
forw(p,f)=∪i=1k−1forw(xi→xi+1,f)=∪i=1k−1SP(xi,xi+1,f)。
S
P
(
x
i
,
x
i
+
1
,
f
)
SP(x_i,x_{i+1},f)
SP(xi,xi+1,f)表示重路由之后两点的最短路径。
关于路径是robustly disjoint的定义：路径
p
⃗
1
\vec p_1
p1与路径
p
⃗
2
\vec p_2
p2所有的边不相交，在路径某些边失效后，重路由之后的路径依然不相交。
优化问题RDP(Min-max robustly disjoint sr-path problem)，找出这样的路径，且两条路径的最大时延最小。
这个问题是NP难的。因为[2]证明了求不相交路径问题是NP-complete的。
关于SR场景的问题搜集
- [16], SR的增量部署，MILP问题
- [18] ,K段问题K-MILP，使用CPLEX求解
- [19] In theory it was shown by [10] that for demand optimization with SR, selecting only one intermediate segment per demand is sufficient. One 2SR optimization took between
80 and 110 seconds. - The authors in [21] introduce an integer linear programming (ILP) model to heuristically evaluate the TE performance of SR-based packet networks. Three different ILP models are proposed, the ECMP model, SHP model, and SEGMR model. Because the computational complexity of the ILP model is high, some instances require too much time to be resolved. Thus, the authors propose a heuristic approach to determine the unique route between each pair of nodes that need to transmit information.
- [22]. The algorithm uses the criticality and congestion index of links to define the link weight. Then, the authors design the Bellman-Ford algorithm with a hop count restriction to solve the problem of minimum weight path.
- [23]使用粒子群算法
[1] Robustly Disjoint Paths with Segment Routing (https://pan.baidu.com/s/1RsfjvCfpQE0TUqZcoKVJpg 提取码: 9xax)
[2] Np-completeness of some edge-disjoint paths problems
[3] Network Flows模型与python代码求解，使用ortools
[4] 网络流（六）最小费用最大流问题
[5] Find maximum number of edge disjoint paths between two vertices
[6] Breadth First Search or BFS for a Graph
[7] Models and Algorithms for Network Optimization
with Segment Routing
[8] IO竞赛 图论—网络流
[9] REPETITA: Repeatable Experiments for Performance Evaluation of Traffic-Engineering Algorithms
[10] Optimized Network Traffic Engineering using Segment Routing
[11] CG4SR: Near Optimal Traffic Engineering for Segment Routing with Column Generation
[12] Expect the Unexpected: Sub-Second Optimization for Segment Routing
[13] A Survey on Replica Server Placement Algorithms for Content Delivery Networks
[14] Node-Constrained Traffic Engineering: Theory and Applications
[15] Centrality-based Middlepoint Selection for Traffic Engineering with Segment Routing
[16] Incremental Deployment of Segment Routing Intoan ISP Network: a Traffic Engineering Perspective
[17] A Scalable and Error-Tolerant Solution for Traffic Matrix Assessment in Hybrid IP/SDN Networks
[18] Traffic Engineering in Segment Routing Using MILP
[19] Traffic Engineering Using Segment Routing and Considering Requirements of a Carrier IP Network
[20] MPLS-based reduction of flow table entries in SDN switches supporting multipath transmission
[21] Traffic engineering in segment routing networks
[22] An efficient routing algorithm based on segment routing in software-defined networking
[23] An Optimization Routing Algorithm Based on Segment Routing in Software-Defined Networks
[24] Novel SDN architecture for smart MPLS Traffic Engineering-DiffServ Aware management
[25] SDN-Based Data Center Networking With Collaboration of Multipath TCP and Segment Routing
609
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
