Essay:

Essay details:

  • Subject area(s): Engineering
  • Price: Free download
  • Published on: 7th September 2019
  • File format: Text
  • Number of pages: 2

Text preview of this essay:

This page is a preview - download the full version of this essay above.

ABSTRACT '

In the Current time every organization design its own intranet when they work on the different distributed infrastructure and connect on one another with the help of intermediate device such as router and switch. To Implementation of infrastructure based on different issue like configuration, maintaining and different topology to design a reliable distributed system. In the paper we design a network with the help of network simulator to check the different structure of intranet and reliability to distribute the load of a network among its various edges link. To design an infrastructure we use different apparatus like switch, router, Ethernet and serial connection, PCs and use different techniques. The developed system is disintegrated to form a fault tree and its reliability is evaluated with the help of fault tree analysis.

KEYWORD: - Router, Switch, Fault tree Analysis, Serial connection, IP Address, Reliable, Ethernet, System reliability  

1. Introduction

In Network based distributed computing has become alternatives to dedicated parallel supercomputer system due to recent advance. We implement these techniques mostly in academics and industries, where we implement computing to benefits the maximum utilization of resource and system [1]. Load balancing algorithm must deal with different unbalancing factors according to applications and to the environment in which it's executed. Distributed system provide cost effective means for extensibility and resource sharing, and obtain potential increase in performance, reliability, and fault tolerance[2],[3]. Distributed system is a collection of several hosts connected by a network where computing share resource between the hosts [4], [5]. An algorithm based on Stochastic Petri Nets (SPN) is presented in [6]. The SPN based approach in uses Minimal file spanning trees (MFSTs) as input to evaluate program reliability. This approach can be easily drawn-out to include execute, and availability. This results in the SPN-based approach being more desirable. Since there is no well-organized arrangement to discover the MFSTs, there is a need to investigate ways of finding the MFSTs for a given distributed system.

Efficient methods to optimize the system reliability have been reported by Raghavendra et. al. [7] which describes reliability of the distributed computing system as dependent not only on reliability of a communication grid but besides on the reliability of the processing nodes and distribution of the resources in the network whereas according to Shatz et al [8], when the system hardware configuration is fixed the system reliability mainly depends on the allocation of resources.

Fault tree analysis is 'To design the system that work correctly we often need to recognize and accurate how they can go wrong' [9]. Fault tree analysis identifies, models and evaluates the unique inter link of events important to failure, undesired events and/or unintended states [10].Performance evaluation of a distributed system based upon fault tree analysis reported by Singh et.al. [11] And whereas according to Bhatia et.al [12] describe the random connection and different form of topology also affect the reliability of the processing nodes and distribution of the resources in the network.

Simulation of a real distributed system assistances in understanding the deeds of it. It supports the learner to improve investigative and design skills by provided that safe and reliable environment for making practice [13]. The current section of research is an effort to design and establish a reliable Distributed System through multiple PC's, routers and switches, and to create them connect with 100% reliability and 0% data or packet damage.

We organize paper in some section as follow. In II section we describe the problem statement Section III is the computational algorithm part where we describe how the simulation model work for to make the data transfer reliable. Section IV is a sample problem of a real distributed system. Section V is for fault tree analysis of problem and we conclude the research work with the future prospects of the concept.  

2. Problem Statement

Let we design a system of a set of distinct nodes are interconnected with each other with the help of different configuration hardware devices like router, switches, and hubs. These small networks established with types of topology and interact each processor with one another with the help of switch. These small distributed networks further connected with different network (using a router) and so on. Researches have to suggest an approach to design and simulate the system so as to make the interaction among each communicating knob as reliable. Further, to compute the genuineness of the system, the structure goes under Fault tree analysis to estimate the reliability of the distributed system.

3. Proposed Method

First of all we start with design a distributed environment is established according to the choice, its design and simulation part activates. The reliability of the system exists in safe allocation of IP addresses, organization of clock rate and demarcation of default gateways. Here we explain the complete design and simulation process with the help of a three ' phase ' protocol.

' Configuration of routers.

' Applying the routing information protocol.

' Assignment of IP addresses and default gateway for the distributed processors.

The computational algorithm for the project and recreation of a distributed system with 'n' processors connected with the help of 'm' routers is as follows:

Algorithm:

1) Input x, y, z, w.

2) Empower the network.

3) Form the connections.

4) For x =1 -- z do

Begin

4.1. Allocate IP address for s0 port.

4.2. Match the clock rate.

4.3. Allocate IP address for e0 port.

4.4. Allocate IP address for s1 port.

4.5. Match the clock rate.

End

5) Run the routing protocol between the links (say RIP).

6) For y =1 -- w do

Begin

6.1. Allocate IP address to the PCy.

6.2. Assign Subnet mask.

6.3. Define default gateway.

End

7) Ping a PC from any other PC in the

Network.

8) If (Loss = 0%) OR (Packets sent =

Packets received)

Print result and exit.

Else

Go to step 4.

Having gone through it, outcome of the complex distributed network where every node is interacting with each other in a reliable manner. Now to evaluate the reliability of these communication channels, the distributed system undergoes fault tree analysis to compute how the reliability configures. If the reliability comes above 0.8, then we can say the system is reliable.

4. Discussion the Algorithm with Problem Inputs

We consider the following distributed system with many processors connected with one another with the help of switches and routers.

In the presented problem (Fig. 1), we have used 4500 series router that contains a 6 Ethernet port, 1 Fast Ethernet port, 6 serial ports and 1Bridge and a 3550 series switch that contains twelve fast Ethernet ports and two Gigabit Ethernet ports.

All the workstations are associated with the aid of these routers and switches to form this distributed environment. Here we connect different router with Ethernet connection and router with switch through

this distributed environment. Here we connect different router with Ethernet connection and router with switch through

 Serial connection.

As per the set of rules, the plan and replication part is done with three ' phase ' protocol. Here is the detailed employment of the algorithm.

PC NAME IP ADDRESS DEFAULT GATEWAY SUBNET MASK

PC 1 11.1.1.3 11.1.1.1 255.0.0.0

PC 2 11.1.1.4 11.1.1.1 255.0.0.0

PC 3 12.1.1.3 12.1.1.1 255.0.0.0

PC 4 12.1.1.4 12.1.1.1 255.0.0.0

PC 5 12.1.1.5 12.1.1.1 255.0.0.0

PC 6 13.1.1.3 13.1.1.1 255.0.0.0

PC 7 13.1.1.4 13.1.1.1 255.0.0.0

PC 8 14.1.1.3 14.1.1.1 255.0.0.0

PC 9 14.1.1.4 14.1.1.1 255.0.0.0

PC 10 15.1.1.3 15.1.1.1 255.0.0.0

PC 11 15.1.1.4 15.1.1.1 255.0.0.0

PC 12 15.1.1.5 15.1.1.1 255.0.0.0

PC 13 16.1.1.3 16.1.1.1 255.0.0.0

PC 14 16.1.1.4 16.1.1.1 255.0.0.0

PC 15 17.1.1.3 17.1.1.1 255.0.0.0

PC 16 17.1.1.4 17.1.1.1 255.0.0.0

PC 17 17.1.1.5 17.1.1.1 255.0.0.0

PC 18 18.1.1.3 18.1.1.1 255.0.0.0

PC 19 18.1.1.4 18.1.1.1 255.0.0.0

 In the first phase, IP addresses are assigned to each of the router for its distinct serial and Ethernet ports.

Step 1: Configuration on the Router

Table 1:- Router Configuration

ROUTER NAME PORT INTERNAL IP DEFAULT GATEWAY

R01 E0/0 11.1.1.1 255.0.0.0

E0/1 10.1.1.1 255.0.0.0

R02 E0/0 12.1.1.1 255.0.0.0

E0/1 10.1.1.2 255.0.0.0

E0/2 20.1.1.1 255.0.0.0

R03 FE0/0 13.1.1.1 255.0.0.0

E1/0 20.1.1.2 255.0.0.0

E1/1 30.1.1.1 255.0.0.0

R04 E0/0 14.1.1.1 255.0.0.0

E0/1 30.1.1.2 255.0.0.0

E0/2 40.1.1.1 255.0.0.0

R05 FE2/0 15.1.1.1 255.0.0.0

E0/2 40.1.1.2 255.0.0.0

E0/0 50.1.1.1 255.0.0.0

R06 FE0/0 16.1.1.1 255.0.0.0

E1/0 50.1.1.2 255.0.0.0

E1/1 60.1.1.1 255.0.0.0

R07 FE1/0 17.1.1.1 255.0.0.0

E0/0 60.1.1.2 255.0.0.0

E0/1 70.1.1.1 255.0.0.0

R08 FE1/0 18.1.1.1 255.0.0.0

E2/0 70.1.1.2 255.0.0.0

Step 2: In the first phase, to configuration of different router here we use Routing Information protocol to assigned IP address within two adjacent routers. The process reprises itself till the all routers have been constructed.   

Step 3: Configuring the Distributed Processors

Different network structure distributed processors are assigned IP address, default gateway and subnet mask as illustrated in Table 2:- Processor IP Address Assignments

When we successfully complete all there phase, here we found the reliable distributed system. To check the working of the algorithm, one can ping any processor from any other processor in distributed system. To show the result of the functionality of the system, processor P17 has been pinged Processor P3 and the communication result is shown in figure .2.

5. Reliability Analysis of the System using Fault Tree Analysis

Fault tree analysis is a top down, rational failure analysis in which an undesired state of a system is scrutinized via Boolean logic to combine a succession of lower-level events. The fault tree explicitly shows all the altered relationships that are obligatory to result in the top event. In constructing the fault tree, a thorough considerate is attained of the logic and basic causes leading to the top event. [17] The fault tree provides a context for all-inclusive qualitative and quantitative valuation of the top event.

FTA is used to resolve the causes of system failure and it's used to quantify system failure probability. FTA is used to evaluate potential upgrades to a system and it's used to elevate resources in declaring system safety. Fault Tree is used to model system failures in risk assessments. [9]

The fault tree can easily be constructed of the above-distributed system as all the nodes are connected in series in figure 3.

To examine the failure rate of the routers, we pinged different node from one another and index their packet loss. After iterating that process on different packet no, we scheduled a table for it and examine the failure rate.

Table 3 :- Communication from nth Processor to (n+i)th Processor

Source Node Destination Node No of Packet Send % of Packet Loss

PC13 PC3 406 1

875 0

PC12 83 1

614 2

PC1 147 0

374 1

PC3

PC7

76 0

405 1

PC15 50 1

211 2

PC10 201 0

90 1

PC 16 PC2 435 1

234 0

PC6 76 1

121 2

PC5 786 0

345 1

During the simulation run on BOSON Netsim [13], a total of 5529 packet have been transferred from one processor to another one to evaluate the reliability of distributed system. Simulation outcomes show that throughout the entire process of communiqu'', a total of 15 data packets have lost out of 5,529 packets that comes out to be packet loss of 0.277%. This packet loss specifies probability of failure in the designed Distributed system and hence disturbs the reliability of the system. This packet loss may be due to numerous reasons viz. network congestion, connection failure etc. Subsequently this packet forfeiture cannot be deterministic on each router and may vary in different runs owed reasons specified above, hence, we have to attain the statistical probability of failure of every single router. Subsequently the packet loss of the premeditated system comes out to be 0.277%, we may use random events method [14] [15] [16] to generate possibility of failure of every router. Accordingly, Table 4 depicts failure rate of routers.

Table 4:- Failure rate of Router

Router No Probability of Failure

Router1(POF1) 0.001

Router2(POF2) 0.003

Router3(POF3) 0.002

Router4(POF4) 0.007

Router5(POF5) 0.004

Router6(POF6) 0.005

Router7(POF7) 0.001

Router8(POF8) 0.003

Now, if failure of the complete distributed

System is given by F(S),[17] then,

F(S) =P [F (1) OR F (2) OR F (3) OR F (4) OR F (5) OR F (6) OR F (7) OR F (8)]

= P [(F1 OR F2)] OR P [F(3)OR F(4)OR F(5)OR F(6)OR F(7)ORF(8)]

= [P (F1) + P (F2) ' P (F1F2)]

                       OR

P[F(3)ORF(4)ORF(5)ORF(6ORF(7)ORF(8)] = [P(F1) + P(F2) ' P(F1F2)] OR [P(F3)+ P(F4) ' P(F3F4)] OR [P(F5) + P(F6) ' P(F5F6)] OR [P(F7) + P(F8) ' P(F7F8)]

= [0.001+ 0.003-(0.001*0.003)] OR [0.002 + 0.007' (0.002*0.007)] OR [0.004+ 0.005-(0.004*0.005)] OR [0.001 +0.003-(0.001*0.003)]

= [0.003997] OR [0.008986] OR        [0.00898] OR [0.003997]

 =0.02596

Now, as per the classification of reliability [15], Reliability of the distributed System R(S) is,

R(S) = 1 ' F(S)   = 1 ' 0.024596

R(S) = 0.97404

Hereafter, the reliability of the configured

Distributed System comes out to be 0.97404.

6. Conclusion

In this paper, an work has been prepared to converse all the particulars about how to form a linking between distributed processors and promote to deliberate the reliability of the system, how to paradigm a block graphic of Fault Tree and how the reliability of the system is assessed through FTA. Distributed computing situation has become needed for all sort of distant infrastructures, but from experience, we cultured that it is not so easy to create, format and manage straight a small area network. It requires a lot of exertions to keep a distributed system up and administration in reliable manner. Simulation has become a very convenient tool in many grounds of routine life and it plays a very important role in education and science. Simulations provide us with the capability to understand and get a good feel about the effects without genuine risk intricate in the given situation. Further, different categories of routers and switches can be used in the establishment of different categories of distributed systems, depending upon the user requirements. Also, different routing protocols such as IGRP, EIGRP, OSPF, BGP etc. can be applied between the connections of the routers depending upon the requirements. Once, the distributed system is configured, we assumed it to be comprised of sub components i.e. routes & decomposed the above system in the form of a fault tree. Assuming the probability of failure of the routers, we then intended the probability of disappointment of the routers established on the concept of Fault tree analysis and finally we succeed in finding out the reliability of the Distributed System that comes out to be 0.97404.

7. Reference

1) Osman , H. Ammar' A Scalable Dynamic Load-Balancing Algorithm for SPMD Applications on a Non-Dedicated Heterogeneous Network of Workstations (HNOW)'

2) N. Lopez-Benitez, Dependability Modeling and Analysis of Distributed Programs, IEEE Trans. Software Engineering, Vol. 20 No 5, pp. 345- 352, May 1994.

3) Kumar, S. Rai and D. P. Agrawal, On Computer Communication Network Reliability under Program Execution Constraints, IEEE Trans. On selected areas in Communications, Vol. 6, No 8, pp. 1393-1400, October 1988.

4) V. K. P. Kumar, S. Hariri, and C. S. Raghavendra, Distributed Program Reliability Analysis, IEEE Trans. On Software Engineering, Vol. SE-12, No 1, pp. 42-50, Jan 1986.

5) M. S Lin and D. J. Chen, General Reduction Methods for the Reliability Analysis of Distributed Computing Systems, The Computer Journal, Vol. 36, No 7, 1993.

6) N. Lopez-Benitez, Dependability Modeling and Analysis of Distributed Programs, IEEE Trans. Software Engineering, Vol. 20 No 5, May 1994, pp. 345-352.

7) Raghavendra, C. S., and Hariri, S., 'Reliability Optimization in the Design of Distributed Systems', IEEE Transactions on Software Engineering, Vol. SE-11, pp. 1184-1193, 1985.

8) Shatz, S. M., and Wang, Jai-Ping, 'Models and Algorithms for Reliability-Oriented Task, Allocation in Redundant Distributed Computer Systems', IEEE Transactions on Reliability, Vol. 38, No. 1, pp. 16-27, 1989.

9) Dan Goldin, NASA Administrator, Fault Tree Analysis, Clifton A. Ericson II, 2000.

10) Sutton, Ian S., Fault Tree Analysis, Sutton Technical Books, Houston, Texas, 2011.

11) Singh Shipra, Garg M. L., Performance evaluation of a distributed system based upon fault tree analysis. IOSR Journal of Computer Engineering (IOSR-JCE) e-ISSN: 2278-0661, p-ISSN: 2278-8727, Volume 17, Issue 1, Ver. IV (Jan ' Feb. 2015), PP 06-12.

12) K. Bhatia, P. K. Yadav, And Sagar Gulati,  Design And Simulation Of A Reliable Distributed System Based On Fault Tree Analysis International Journal Of Advanced Computer And Mathematical Sciences Issn 2230-9624.Vol 2, Issue 3, 2011, Pp 168-175 Http://Bipublication.Com

13) Boson Netsim 6.0: http://www.boson.com/Product/CIS-NS-CCNP-02.html

14) L. S. Srinath, Reliability Engineering, Fourth Edition, East west press, 2006.

15) P. Jalote, An integrated approach to Software Engineering, Narosa Publishing House, Second Edition, 2003.

16) Aggarwal, K., K., and Yogesh Singh, Software Engineering, New Age International Publishers, Revised Second Edition, 2005.

17) Bill Vesely,'Fault Tree Analysis (FTA): Concepts and Applications' Mission Success Starts With Safety.

...(download the rest of the essay above)

About this essay:

This essay was submitted to us by a student in order to help you with your studies.

If you use part of this page in your own work, you need to provide a citation, as follows:

Essay Sauce, . Available from:< https://www.essaysauce.com/essays/engineering/essay-2016-06-15-000BHW.php > [Accessed 20.10.19].