Since the goal of this project is to develop and apply a fault tolerance system for sophisticated distribution systems with the help of mediator based approach. The entire process is clearly explained in this chapter. Here we use mobile agent as a mediator and hence it is known as mobile agent based approach. It is used to find the failures around the network systems and fault tolerant system is also given across the application. Here everything including both the front end and backend design is explained below in detail.
Front end design:
With the help of Java an imaginary dynamic distributed network is developed and this is supported by two servers and also four client modules and the nodes linked with these are expected as distributed in environment. Each sever cannot handle more than two clients hence required two servers to maintain four clients .These particular servers are again maintained by one master sever and these are maintained in such a way that they are distributed in single master server. These two servers act as demo of the master server. The entire scenario including design as well as programming is explained below.
Master server
Master server plays a very crucial role in this particular scenario. As this is required for the maintenance of the other two severs. These two servers do not depend on each other and are well distributed. Master server is shown with the help of icon and this is used by end-user makes the process to begin. Start and stop modes are provided to the clients once the start mode is pressed the master server initiates the two servers. When the users choose the “stop” mode is enabled through any of the servers the master sever automatically gets switched off .The starting process is explained in detail this process is given below.
Server modules:
As we discussed two servers are developed and acts as distributed servers and these servers can hold two clients and in upcoming many clients can be attached to it while designing taking into considerations and this facility is given ate developing level itself. A sever module consists of few components which is briefed in later stages. And this server is named as sever A while other one is named as server B ON and OFF buttons are given for each server this is done with the help of server icon once the on button is enabled the server gets started and these operations are performed as given below.
Send data
The initial operation and the important operation available is to send the information and his can be done by simply clicking the button. Once the client clicks this button a pop up window opens and asks for the file to select as the file is selected the further step is to choose the client since each server can take-up two clients, these clients can be identified with IP address or client name. A ideal number is allotted to each client for identification this is known as IP number or IP Address of the particular client as the application is outstanding it cannot be demonstrated through N/W operations.
Once the process is done by selecting the client automatically the sending process starts and this can be viewed graphically till the information reaches the client. Here one more option is provided to user the total number of packets transferred to client can be viewed. In general the users can look how many data packets to be transferred and how many data packets can be received at the other end user can transmit these to the selected client as discussed earlier and the data sent is saved in the client’s back end with the server name and records its data packets and the client id.
Receive data
Servers gets the data transferred from clients to receive this another button is supplied to the users in the server icon .Once this is done the server receives the information from the particular client which contains the information and the data received is saved in server i.e. in back end we can count number of packets also. For this to happened the communication link should be perfect this is explained below
Link status
Link status acts as a scavenger in this process it is connected between server to client and the status is obtained from both the ends before sending it as well as receiving it once the status is ON, to know the status of the data whether received successfully or not and the off status denotes that data sent is not received and failures are given. This can be done at any point of time and if any failure occurs the status default turns to ‘OFF’ from this we can access that failures exists in the system around the network and once the failures are picked based on this a fault tolerant system is designed and by implementing it the link status is set to on for communication process
Add clients
Adding of clients is also done the users this is done by the button provided in the server icon once the client is additional to the network an IP address is given to the client and everything is processed by this number. The addition of client leads to create a row in backend and all the details are registered in this row the server icon contains a add click button this gives a option to select the required server for which clients need to be appended. Adding of clients is very sensitive because each sever supports only 6 clients hence this done carefully.
Distributed network module:
For this application to be considered as distributed network one distributed network is shaped individually some assumptions are done in this context. In this scenario a metrics is created by fixing the clients required and the data to be transmitted and if this parameters which are chosen crosses the range a backup server is selected for addition of clients and this is utilized to maintain the requests the prerequisites to be maintained in this scenario are given below
- A server maintains only 6 clients if more than that occurs it is automatically appended to the other server.
- An extra server is further created depending upon the client’s requirement once the 6clients limit is reached.
- As we know each server definitely handles the 6 client requests for to and fro of the data once it across the boundary in this scenario.
- Master server maintains all the updates in the level sever as discussed above
Client module:
A brief explanation to the network system (distributed) and the server scenarios are explained better. These designs are delivered for the client module is explained here. As we all know a server can maintain 6 clients these clients will intercommunicate with each other or they can also communicates from one server to other as the limit is exceed following are the important operations of the client module.
Send data
Clients can pass the data to the required server or to the successive client depending upon the strategy this strategy is designed by the end user whether to connect with the server or the successive client. Naturally the users are provided with some choices and one of them is peer clients with the particular server and in any way the distributed network circumstances are ruined a warning message appears that the sever is not free and now the servers can select another one to send the particular request. While passing the data users are given the option to choose the field it maybe text field once the selection is over send button is pressed and these details are automatically saved in the back end. The data sent may be gathered by server or other client within he server or client of the other sever.
Receive data
Clients gather data either from the server or client and for this purpose a particular module is designed for this scenario. A link is available to check the data whether received from the server or client. Users are provided with a facility as they can click on the link if required a table pop ups showing all the particulars of the received from the sender and the columns in the table are sender id, time and date in which data is sent and also the content of the data once the entire process is done the user can download all the details when required. This data can be easily fetched from back end for identifying the faults as given below.
Fault detection module
The main goal of this project is to access a fault tolerance system using mobile as agent for this particular system a fault detection system is introduced in this scenario. To identify this faults few of the errors are introduced knowingly and once this is done this is invoked to identify the corresponding faults and these faults can be rectified by using this agent. These are explained in detail
Fail the link
The communication links between client and server are made to fail in this scenario this can be obtained by the end users by selecting the appropriate link in medium the server and the corresponding clients. For this purpose a separate button is given to the users once the link is clicked it is intentionally made to fail when the user selects it particularly .The status automatically turns to OFF and this is seen in the status scenario and at the same time it is appended in the status log also. A Failure table is also created and all the log details are maintained here. This is explained with better example with link failures and count of nodes between the link and the failure time occurrence is noticed in the table
Distributed network failures
When the users opts this one the scenarios which are discussed earlier does not work anymore and also makes the other servers or clients not to work to send the requests. the status automatically turns to off and once this is done distributed network option no more works and all the details are recorded in the failure database. This is explained by a simple scenario two simple failures are taken into consideration in this process .once the user detects these failures a special button is given once this pressed all the failure steps are viewed by the end users and these particulars are collected from the failure backend system.
Mobile agent based fault tolerance module
Here comes very important part i.e. mobile agent based fault tolerance to identify all the failures and to come with a better solution for these failures. For this agent selected is mobile agent this fetches all the failure results from the database table and starts operating on them. This is started by pressing the run button and after detection it delivers solution this are explained further.
The link failures are easily detected if they occur here this agent will approach other parallel links this can be understand by example suppose there is link failure from client to server B or other client this mobile agent does it works by scanning it and proposes other links to the server or client sometimes both bypassing through other clients .the users always has best option to choose with the help of this mobile agent And the link status is set to ON the users can move on with this mutual transmission.
The other failure is distributed network failure when the user doesn’t make utilize of this choice. The particular mobile agent repairs it the particular scenario in the discussed manner. Mobile agents work is to add more servers to this request and clients are originated based on these servers provisionally and make sure the network option is ready to work and disconnects once the actual fault is cleared and the distributed network is enabled and this choice is made regularly whenever the problem occurs above discussed manner.
Paper Submitted & Written by Sathish Nagarajan