RILCE:
An approach for Resource Indexing in Large Connected Environments

I. Approach

In our approach, we proposed a decentralized approach for indexing resources in an open environment. An index is created by a super node named orchestrator. The orchestrator has a vision of the entire environment. The super node creates an index having the devices as rows and zones as columns. This index contains bits representing the index knowledge of each device. These bits are filled using the indexing algorithm explained in our paper. Our algorithm takes into consideration the devices' capacities. We note that α and β are two variables used by the algorithm to determine the number of indexes in covered and uncovered zones. The first step of this algorithm is to assign each device to its zone. The second step is to build a minimal cycle between them. The third step is to create a diagonal on the uncovered zones. In this step, if devices still didn't receive an index during this step, indexes are added in the covered zones part. After this step completion, we create an alternation between the addition of indexes in the covered zones and in the uncovered zones until the algorithm converges (α = 0 and β = 0 or all devices are filled till their maximal capacity). The different algorithm steps are shown below along with the system architecture in figure 1 and figure 2. Then, we can start sending queries created by our query generator.

Fig. 1. Simulator tool snapshot	Fig. 2. Simulator tool snapshot

II. Query Generator

In order to support our work, we decided to show a glips of our work by showing our query generator. We will also explain the different steps along with its result.

The following generator could be downloaded on this link : https://sigappfr.acm.org/Projects/RILCE/generate_queries.exe

Input the covered zones seperated by a comma(z0,z1,z2,z3,z4,z5 in the attached environment): z0,z1,z2,z3,z4,z5

Input the uncovered zones seperated by a comma(z6,z7,z8,z9 in the attached environment): z6,z7,z8,z9

Input the number of queries needed: 20

Input the percentage of queries going to a covered zone: 0.7

Select the generated environment path(attached environment example)

We can see in the part above the way to generate the queries using our generator.

Note that in the last step the attached environment-example.csv must be selected.

The output must be 2 csv files. The One contains the actual queries while the other csv contains settings about the envrionement used by the generate index algorithm