Enhancing Data Exchange with Peer-to-Peer Integration Systems

The Role of Schema-Based Peer-to-Peer Integration Systems

A. INTRODUCTION

CE 1.1

This 1^st part presented here gives the explanation of the activities that I performed in the project entitled “Implementation of Schema-Based Peer to Peer Data Integration System”, which I undertook as an academic requirement while I was pursuing B. TECH in Electronics and Telecommunication Engineering from Jawaharlal Nehru Technological University. It is located at Hyderabad. I was supervised by the project guide and I was assigned project leader for this project. I did this project during the academic year …….

B. BACKGROUND

CE 1.2

P2P networks when used for a distributed information system, all the peers are allowed for information access and providing information, unlike the centrally organized system. Seeing the loose and dynamic association amongst peers in P2P integration meaning it is based on files exchange identified through attributes of a limited set and without using schemas for data description, this project was undertaken to demonstrate the BAV (both-as-view) mechanism for P2P integration at the schema level. The integration is achieved between the schemas when both of the schemas have a pathway to some of the published schemas. The message exchange between peers is shown through data integration system in this project.

CE 1.3

The project focused on developing a data integration system to a peer-to-peer network, for which I considered a few objectives that are listed below:

• 
  To develop and configure a peer to peer network for achieving integrity
  
• 
  To optimize peer-to-peer network so that it could provide gateways to the different public schema
  
• 
  To create the BAV pathways for sending of the update requests for changes in peer scheme
  

CE 1.4

As a project leader, it was my duty to see the managerial aspects of the project, which involved the documentation work like preparation of report, preparation of presentation slides, giving the presentation, dividing the work in the team, collaborate with the guide for providing status updates, come up with ways for completing the project tasks within the scheduled time. And, I worked on the design and development of the configuration of different peer-to-peer networks which could achieve integrity. I also implemented modular communication between peers to achieve targeted goals. I coded program in AutoMed for creation of objects and integration of data.

CE 1.5

The chart provided below explains the hierarchical structure I managed to follow for the execution of the project. It also includes my exact role for this project.

Figure: Organizational Chart

CE 1.6

The list of duties I executed besides the managerial duties for the project is summarized below in abridged points:

• 
  To design the prototype of peer to peer structure
  
• 
  To generate the pathway for data transmission
  
• 
  To develop and implement the infrastructures for achieving the modular communication between the peer.
  
• 
  To implement serialization of the AutoMed objects for achieving the data transmission.
  
• 
  To develop a program for handling the object creation.
  
• 
  To conduct tests on the designed network with the use of three peers.
  

C. PERSONAL ENGINEERING ACTIVITIES

CE 1.7

I, first, went through the literature study where I reviewed different existing papers on a peer-to-peer network, different integration types, and data management in P2P networks. I used AutoMed software for the development of the data integration system for the P2P network. I used super-peer schemas as using global schema without considering physical super-peer nodes. I developed the configuration for the peer to peer network for achieving integrity. I made sure that the registration process was running within the same host for achieving the perfect operation of the peer. I also specified a port number for the registration purpose.

CE 1.8

The project demanded the application of sound knowledge on electronics and telecommunication engineering. I applied my knowledge and understanding regarding network management, data information system, real-time operating systems, communication systems, etc. for the execution of the project at various phases. I followed IEEE 802.11 and IEEE 802.15 standards as this project used wireless LAN and wireless PAN as a peer to peer network.

CE 1.9

• 
  After the review, I collected the software required for the project: AutoMed, which is a software package that consists of graphical tools and APIs, providing a solution to the data integration problem. Here, I used IQL (intermediate query language) for example queries, which in AutoMed system is used by the exact transformation of schemas. I used UDDI directory services for peers publishing which is integrated using standard super-peer schemas. In order to integrate the ps_i peer schema with SPS super-peer schema, I firstly generated a pathway for ps_i → SPS.
  
• 
  I found the SPS construct that had no relationship with peer schema and added those constructs using [Void, any] extent for extending transformations. Now, for the constructs that remained (c) in SPS, I used extend transformations on psi so to find construct c with sound query specifying constructs in psi used for deriving construct’s global instances. Then, I decided which psi constructs had no relationship with super-peer schema, and then used contract transformations for removal of such constructs with use of [Void, Any] extent.
  
• 
  Then, I used these pathways for updating the request. For this, I performed series of steps that performed message translation in two-phase: logical translation to make sure that the update request from data source is applicable to the data source where it is to be applied, and operational interpretation of the message for decision making of whether the update be applied to data source or not. I defined the super-peer min-max interpretation in order to send data from one peer schema to another through super-peer schema, where minimal and maximal interpretations were taken for deriving construction extent in super-peer schema from peer schema one and peer schema next from super-peer schema respectively.
  
• 
  Then, the query requests were sent over BAV pathways. Here, I developed and implemented infrastructures for achieving modular communication between the peers. I designed the prototype of the peer to peer schemes for achieving data integration. I coded a program for the creation of the AutoMed objects with the use of the AutoMed tree. I made use of Abstract Daemon for achieving multithreading process with the capability of binding the ports and listening to request. I defined a class with the possibility of reusing and reconfiguring. I also made use of the abstract server as the subclass of the abstract daemon for handling the request targeted to Abstract Daemon via an interface for request listening. I also implemented the registry of the abstract server.
  
• 
  I performed the serialization of the AutoMed objects for achieving the data transmission from one AutoMed repository to all other AutoMed repository connected within the network. I made the selection of the serialized AutoMed objects with an XML format. I developed the architecture for achieving the flexible representation of the schema of the AutoMed object using the XML format. I made the conversion of the Document to XML and vice versa by performing the serialization using Xerces API. I worked on the conversion of the XML file into document form with the use of Javax.xml.parsers package of the Java API.
  
• 
  I coded the program for handling the object creation. I made the use of the object builder class for the creation of the appropriate objects. I was involved in the creation of the serialization and realization of the AutoMed objects. I had developed the intermediate representation rather than XML format for providing the flexibility for the selection of various formats of the object. After the completion of the development of targeted peer to peer network for the data integration, I tested the designed network with the use of three peers. I posted the queries from the different peers to the schema and analyzed the generated outcomes. I found that the network was not affected while sending the same broker query via the schema.
  

CE 1.10

TECHNICAL PROBLEMS AND SOLUTIONS

• 
  One of the issues, I was faced in the project was regarding the renaming and deletion of the object created with AutoMed. I encountered the problem while I was implementing AutoMed i.e. updating the created AutoMed objects using the AutoMed tree, where the system was displaying the alert that the object couldn’t be deleted when I was updating the object’s name. While solving the problem, I deleted the object to be renamed and created the new object with the name required by our system. But this idea failed where I was updating and deleting the objects at a nodal position, as here all the objects connected with the nodal objects was also being deleted and updated. I was lagging behind the estimated time of the project due to the generated problem. So, I brought up the problem to my supervisor, who suggested me to apply the rename transformation to the schema objects made up of the special type of array. The array was defined to record all the objects that were deleted and updated again. I performed the rename transformation for the updating, creation, and deletion of AutoMed java objects. I defined a special array capable of encapsulating all the deleted and updated objects. Finally, I performed the rename operation on the created objects as per the system requirements.
  

• 
  Another issue I encountered in the project was about the representation of the pathway. During the development of a pathway for the list of transformations after defining the peer with the client and server and using the AutoMed java objects with the use of the AutoMed tree, I was made responsible for the recreation of the pathway in one section where some problems were created earlier. I used the transformation order used in the earlier pathway for the creation of the newer one but the pathway was not created. So, I consulted my senior and gained an idea of both schemes ident presence requirement during the time of transformation, which I implemented. I created the subnet for both transformations and then performed the ident transformation. I made the use of an abstract network for pathway representation with the purpose of building the transformations in the proper order. I also programmed for implementation of the naming scheme and command was given for the use of the alternative names when the problem of naming conflict was raised. Finally, with the use of the subnets and abstract network the correct pathway for the peer to peer data integration system was built.
  
• 
  Another problem I encountered was regarding the addition and movement of the peer. During the testing and simulation of the system for performance analysis, when I added a peer in one of the repositories of the system for observation of the system flexibility and conducted simulation, I found that the graphical editor of the AutoMed was not responding with the changes made on the system. Then, I removed the added peer and simulated the system for performance analysis; I observed the desired result of the system during the simulation. I discussed the encountered problem with my team members who suggested me to restart the editor for the adaptation of the changes made in the repository. I restarted the editor of the system and found that the refreshing of the object cache was made for the adaptation of changes made in the repository. I also faced the problem during the movement of the peer within the network for the simulation purpose. The system was not identifying the peer. To solve this problem, I introduced the peer description by addition of the peer name and its specific IP address. Then, I simulated the system; I found the acceptable output. I also repeated the test of the system by changing the peer location. I found the system was responding with the peer movement and desired output of the system was also achieved.
  

CE 1.11

CREATIVE WORK

Here, I used a BAV (both-as-view) approach for data integration rather than going with alternative approaches like GAV (global-as-view) and LAV (local-as-view). As a result, a richer integration framework was provided for data integration, where a complete mapping between peer schemas was possible in both direction, and so proved very useful for the p2p application. Besides this, there were many works that could count as innovative, like I defined and implemented the abstract network for achieving the perfect pathway between peers. During testing, I added a peer in one of the repositories of the system for observation of the system flexibility.

CE 1.12

I saw all of the managerial activities during the conduction of this project as I was the project leader. I divided the work among the team and gave different assignments to the associate members of the project. I used different project management tools like PERT chart and Gantt chart to achieve the project target within the specified time. I organized meetings where I discussed the issues related to the project. Furthermore, I collected various design data and simulation result and used them while developing the project report. I also conducted the project presentation to the department members including the head of the department. I prepared the report and submitted to my guide for review, and in between continuously provided the progress report of the project.

D. SUMMARY

CE 1.13

The schema-based peer to peer data integration system was developed and implemented successfully with acceptable simulation results in this project. The use of BAV approach for data integration system was shown successfully, with a demonstration of queries in transformations, data movement between peer schemas, updating of the path and reflection of the changes in the system.

CE 1.14

I used the both-as-view approach of data integration system for peer to peer system, where I designed the prototype of the peer schema and developed the data integration system using the AutoMed software, where I coded the program to create the objects with the use of AutoMed tree. I created the pathway for sending of the update requests using transformations. I then made changes to the peer by sending update requests during the simulation.

CE 1.15

I acquired a good deal of technical knowledge regarding the P2P networks and data management in the P2P network, different approaches for data integration system and its development. My coding skill was developed, and I got acquainted with the AutoMed software. Besides the technical knowledge I was favored with, I got to work on my managerial skill in this project, and thus developing my coordination skill, management skill, presentation skill, communication skill, etc.