nsp4j-optimizer

Network and Service management Planning framework for Java - Optimizer

Prerequisites

• Apache Maven
• openjdk
• Gurobi optimizer

Build and run

1. Install gurobi.jar library on your local maven repository:

   mvn install:install-file -Dfile=/usr/share/java/gurobi/gurobi.jar -DgroupId=com.gurobi -DartifactId=gurobi-optimizer -Dversion=10.0.0 -Dpackaging=jar

2. Build:

   git clone https://github.com/fcarp10/nsp4j-optimizer.git
   cd nsp4j-optimizer/
   mvn install
   mvn package

3. Run:

   cp src/main/resources/scenarios/{example.dgs,example.txt,example.yml} target
   cd target/
   java -jar nsp4j-optimizer-${VERSION}-jar-with-dependencies.jar

4. By default, GUI runs in: localhost:8082.

5. Check in target/results/ for the generated results.

API

Upload a file:

curl -s -X POST http://SERVER:PORT/upload -F 'file=@test.txt'

Delete a file:

curl -s -X DELETE http://localhost:8082/delete/test.txt

Run a model (example):

curl -s -X POST http://localhost:8082/run -d '{ 
  "inputFileName":"example",
  "objFunc":"UTIL_COSTS",
  "maximization":false,
  "name":"LP",
  "constraints":{
    "RP1":true,
    "RP2":true,
    "PF1":true,
    "PF2":true,
    "PF3":true,
    "FD1":true,
    "FD2":true,
    "FD3":true,
    "sync_traffic":false,
    "max_serv_delay":false,
    "cloud_only":false,
    "edge_only":false,
    "single_path":false,
    "set_init_plc":false,
    "force_src_dst":false,
    "const_rep":false
    }
  }'

Stop the optimizer:

curl -s -X GET http://localhost:8082/stop 

Retrieve results:

curl -s -X GET http://localhost:8082/results 

Configuration files

• *.dgs: topology description file
• *.txt: paths file
• *.yml: parameters file

Topology description file

The network topology is specified using GraphStream guidelines:

DGS004
test 0 0

an n1 x:100 y:150 num_servers:1 server_capacity:1000 
an n2 x:150 y:100
an n3 x:200 y:100

ae n1n2 n1 > n2 capacity:1000
ae n2n3 n2 > n3

• an adds a node. The command is followed by a unique node identifier, that can be a single word or a string delimited by the double quote character. Values x and y on the server represent the coordinates of the nodes. For each node, additional parameters can be specified, for instance, the number of servers or capacity.

• ae adds an link. This command must be followed by a unique identifier of the link, following with the identifiers of two connecting nodes. For each link, additional parameters can be specified, for instance, the link capacity.

For further information, see Graphstream documentation.

Paths file

File containing all admissible paths:

[n1, n2, n3, n7, n9]
[n1, n4, n5, n3, n7, n9]
[n1, n4, n5, n6, n7, n9]

Parameters file

Parameters for the optimization model:

# optimization parameters
gap: 0
weights: [0, 1.0, 0]
# auxiliary parameters
aux: {
  "iterations": 1000,
  "offset_results": 0,
  "scaling_x": 1.0,
  "scaling_y": 1.0
}
# service definitions
serviceChains:
- id: 0
  chain: [0, 1, 2]
  attributes: {
    "minPaths": 1,
    "maxPaths": 2
  }
# function definitions
functionTypes:
- type: 0
  attributes: {
    "replicable": false,
    "load": 1.0,
    "overhead": 10,
    "sync_load": 0.1,
    "delay": 10
  }
- type: 1
  attributes: {
    "replicable": true,
    "load": 1.0,
    "overhead": 10,
    "sync_load": 0.1,
    "delay": 10
  }
# traffic flow definitions
traffic_flows:
  - src: "n1"
    dst: "n2"
    demands_specific: [3, 35, 17]
    services: [1]
    service_length: [3]