Description: This undergraduate course aims to provide the students the fundamentals of network management and network planning. Net2Plan is used as a tool to assist the network planning part of the course. Within this part, students learn how to create and use algorithms, for solving network planning problems, with special focus on planning networks with recovery schemes (network protection / restoration). First, students learn some fundamentals on the theory of problem and algorithm complexity. Then, students learn how to design heuristic algorithms for classical network planning and traffic engineering problems, which have been shown to be NP-complete: i.e. node location, topology design, planning of protection and restoration routes, OSPF weight engineering. Heuristic techniques reviewed are: local search algorithms, greedy algorithms, simulated annealing, tabu search, evolutionary algorithms, GRASP and ant-colony optimization algorithms. With Net2Plan, students are able to fast prototype, test and compare their algorithm proposals.
Year: 3
Credits: 7.5 ECTS credits
Degree: Grado en IngenierÃa Telemática (4 years, 240 ECTS credits)
University: Universidad Politécnica de Cartagena
Full-course details can be found here.
- 2018-19. IGP-metric vs. MPLS-TE based
traffic engineering assessment with Net2PlanThis lab work uses Net2Plan to analyze the merits of different strategic traffic engineering alternatives: OSPF with static link costs (in different forms), dynamic MPLS-TE tunnels, and 1+1 Fast Reroute tunnels. This includes dimensioning the required link capacities in each case for given fault tolerance & latency SLAs. Also, uses Net2Plan to analyze the robustness of OSPF with static link costs, OSPF with dynamic link costs, and the same MPLS-TE flavors as before, for traffic fluctuations.
Author(s): Pablo Pavon-Marino
File(s): Wording
- 2018-19. Network analysis with Net2PlanThis lab work uses Net2Plan to analyzo cost-performance trade-offs in an IP network running OSPF protocol. Performances observed will be the per-flow end-to-end latencies, blocking and congestion situations, under different failures and traffic shifts.
Author(s): Pablo Pavon-Marino
File(s): Wording
- Lab. work 1: Topology assignment problems using local search and greedy algorithmsThis laboratory is focused on the development of two algorithms: (i) a node location problem based on a local search heuristic, and (ii) a network ring-generator based on a greedy heuristic.
Author(s): Pablo Pavon-Marino
File(s): Wording
- Lab. work 2: Flow assignment problems using simulated annealingThis laboratory is focused on the development of a non-bifurcated flow allocation algorithm based on the simulated annealing heuristic.
Author(s): Pablo Pavon-Marino
File(s): Wording, Code template
- Lab. work 3: Flow assignment problems with 1+1 protection using tabu searchThis laboratory is focused on the development of a non-bifurcated flow allocation algorithm, including dedicated path protection, based on the tabu search heuristic.
Author(s): Pablo Pavon-Marino
File(s): Wording, Code template
- Lab. work 4 (2013): Flow assignment problems with pre-planned path restoration using evolutionary algorithmsThis laboratory is focused on the development of a non-bifurcated flow allocation algorithm, including pre-planned dedicated path restoration, based on evolutionary algorithms.
Author(s): Pablo Pavon-Marino
File(s): Wording, Code template (1), Code template (2)
- Lab. work 4 (2014): Evolutionary algorithms for the design of an OSPF routing, minimizing congestion under single-SRG failuresThis laboratory is focused on the development of an evolutionary algorithm deciding on the OSPF weights in an IP network, minimizing the network congestion in the presence of single-SRG failures.
Author(s): Pablo Pavon-Marino
File(s): Wording, Code template (1), Code template (2)
- Lab. work 5: GRASP and ACO algorithms for the Travelling Salesman ProblemThis laboratory is focused on the development of GRASP (Greedy-Randomized Adaptive Search Procedure) algorithm and an ACO (Ant Colony Optimization) algorithm, for the Travelling Salesman Problem (TSP)
Author(s): Pablo Pavon-Marino
File(s): Wording, Code template (1), Code template (2)
- Lab. work 6: Projected gradient algorithms for the bandwidth sharing problemThis laboratory is focused on the development of different variants of the projected gradient algorithm, for a bandwidth sharing problem. Algorithm convergence and the possibility of a distributed implementation are discussed
Author(s): Pablo Pavon-Marino
File(s): Wording, Code template