Optimized Distributed Subgraph Matching Algorithm Based on Partition Replication

Round 1

Reviewer 1 Report

The study proposes a distributed subgraph matching algorithm based on Partition Replica (PR-Match) to process the partition and storage of large scale data graph. They propose a heuristic rule based on prediction cost to select the optimal merging plan. Subgraph matching is a very important research topic in the graph theory, which can help the users to extract valuable information from the graph data sets. The study shows an extensive and up-to-date biography on the subject and the methodology used is appropriate and well explained. Some corrections need to be done: Line 127: change “invloving” by “involving”; Line 147: let an empty line before and after Algorithm 1: Graph Data Partition Algorithm; Line 250:   let an empty line before and after   Algorithm 3: Subquery Matching

 

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This paper proposed a 4-step algorithm for subgraph matching on labeled undirected graph. This problem is a hot topic in graph theory. The method is somewhat novel and illustrated comprehensively. Algorithms for different steps are proposed and complexity analyzed. My comments are as follows.

1. Is this the first time partition replica was employed to solve subgraph matching? More reasoning on why this method is chosen should be provided.

2. Real world application of subgraph matching should be added in Introduction so as to highlight the importance of this work.

3. The symbols in Algorithm 1 should be explained more, e.g. hash(v)modk.

4. In Section 3, there should be a definition of the problem in formal format, i.e. start with boldface Problem.

5. The definition of “large graph” (in what sense?) considered in this work should be clarified. Does “large” imply |V| is large or |E| is large or both? This may change the complexity analysis of proposed algorithms. For instance, if |E| is large – a dense network, then the complexity of Algorithm 1 is O(|E|) or O(|V|^2).

6. In line 185, the author mentioned splitting a query graph into multiple start graphs is NP-hard. It might be good to add a reference here.

7. A better proofreading of the manuscript is in need. I list a few grammar issues below:

line 35: lacks of → lack

line 53: difference → different

line 56-59: At the same time … data matching → please brake down and refactor this sentence

line 74: can be find → found

line 76: gInex(misspelled?) → it

line 100: splited → split

line 101: then merged the sub … → fix grammar accordingly

line 116: is isomorphism of → is isomorphic to

line 204: an extra closing bracket

line 212: is appled to → is applied to

line 323: The proposed … algorithm is conducted → is the word “experiment” missing here?

Line 402: stratege → strategy

Author Response

Please see the attachment.

Author Response File: Author Response.pdf