===========
Data Format
===========

Repetita uses a straightforward format for its topology and demands data.

---------------
Topology Format
---------------
The topology format describes nodes and directed edges of the topology.
Each node has a label and a geographical position,
each edge has a label, a source node, a destination node, an IGP weight,
a bandwidth, and a delay.

The node geographical positions can be useful for visualization purposes.
The edge bandwidth can be in any unit,
since the usage depends on the corresponding demands file(s);
we use kilobits per second.
Similarly, we use delays in microseconds.

For instance, file Arpanet196912.graph looks like:

::
   
   NODES 4
   label x y
   0_SRI -122.18219 37.45383
   1_USCB -119.69819 34.42083
   2_UCLA -118.24368 34.05223
   3_UTAH -111.89105 40.76078

   EDGES 8
   label src dest weight bw delay
   edge_0 0 1 1 1000000 678
   edge_1 1 0 1 1000000 678
   edge_2 0 2 1 1000000 869
   edge_3 2 0 1 1000000 869
   edge_4 0 3 1 1000000 1604
   edge_5 3 0 1 1000000 1604
   edge_6 1 2 1 1000000 237
   edge_7 2 1 1 1000000 237

The first line is `NODES` followed by the number of nodes,
then a 1-line memo for a human editor,
one line per node,
one empty line
a line containing `EDGES` followed by the number of edges,
a memo line,
and one line per edge.
There might be several edges with the same source, destination pair.

--------------
Demands Format
--------------
A demands file is associated to a topology.
It contains a list of demands,
that each have a name, source node, destination node and a bandwidth.

For instance, a demands file for the above Arpanet graph might look like:

::

   DEMANDS 12
   label src dest bw
   demand_0 0 1 435117
   demand_1 0 2 532282
   demand_2 0 3 170697
   demand_3 1 0 598778
   demand_4 1 2 265255
   demand_5 1 3 85065
   demand_6 2 0 977314
   demand_7 2 1 353915
   demand_8 2 3 138841
   demand_9 3 0 184807
   demand_10 3 1 66923
   demand_11 3 2 81868

The first line is `DEMANDS` followed by the number of demands,
then a memo line,
and one line per demand.
Notice that source and destination are the number of the node in the
topology's list, so as to not force the node names to be different.
There might be several demands with the same source, destination pair.