ERR Broadcast Scheduling Algorithm

Transcription

1 ERR Broadcast Scheduling Algorithm Sukumar Babu Bandarupalli 1 Neelima Priyanka Nutulapati 2 Dr. P. Suresh Varma 3 Vijaya Institute Of Technology For Women,Vijayawada, JNTUK SRK Institute Of Technology, Vijayawada, JNTUK Adikavi Nannaya University, Rajahmundry sukumar.vitw@gmail.com priyanka.nutulapati@gmail.com sukumar.vitw@gmail.com Abstract Efficient Round Robin Algorithm is an online algorithm for broadcast scheduling in a client server technology. In this paper the ERR is a pull - based broadcast model, where the server stores n unit sized pages (or n different sized pages) of information. This server is connected with multiple numbers of clients requesting for the pages of the server. The server processes the request and sends the response to all the pages that are requesting for the same page at a time. The main use of this broadcast algorithms is instead of processing the pages individually, the ERR algorithm process one request of same type and sends the response to all the clients who are requesting for the same page at that instant of time. This method reduces the overhead on the web container of a server. This ERR algorithms works better than Longest Wait First For Broadcast Scheduling, FCFS, SJF, Priority and Round Robin algorithms. The main advantage of this algorithm is, ERR minimizes the average flow time between the request and response, minimizes the waiting time and turnaround time of the requests. 1. Introduction Generally in client server technology the client sends the requests to the server for a particular page to access. The process each request of the clients and sends the response to the client. The client requests for n-pages available at the server. In this process, it very tedious and becomes overburden for the server in processing each request received from the multiple clients. For processing and sending requests to the clients the web container of the server uses different algorithms, depending on the interest of the vendors. The main use of scheduling id to process all the requests by minimizing the flow time between the client and the server, and to decrease the waiting time of the clients request, and minimizing the turnaround time of the process etc. There are different scheduling algorithms like First Come First Serve, Shortest Job First, Priority, Round robin etc. In broadcasting scheduling algorithms like Longest Wait First for Broadcast Scheduling etc are used. In this paper we defined a new algorithm called Efficient Round Robin Algorithm, which works efficiently and gives optimal results while processing the client s requests. The algorithm proposed in this article is pre-emptive and non-preemptive in nature and attempts to give fair CPU execution time by focusing on average waiting time and average turnaround time of a process. There are different scheduling algorithms with different characteristics which decide selection of process using different criteria for execution by CPU. The criteria for a good scheduling algorithm depend, among others, on the following measures [1]. CPU Utilization: It is the average fraction of time, during which the processor is busy [7,8]. Throughput: It refers to the amount of work completed in a unit of time. The number of processes the system can execute in a period of 46

2 time. The higher the number, the more work is done by the system [7]. Waiting Time: The average period of time a process spends waiting. Waiting time may be expressed as turnaround time less the actual execution time [7]. Turnaround time: The interval from the time of submission of a process to the time of completion is the turnaround time [7]. Response time: Response time is the time from submission of a request until the first response is produced [7]. Priority: give preferential treatment to processes with higher priorities [7]. Fairness: Avoid the process from starvation. All the processes must be given equal opportunity to execute [7]. 2. OVERVIEW OF EXISTING CPU SCHEDULING ALGORITHMS First Come First Served (FCFS) Scheduling. It is the simplest CPU Scheduling algorithm. The criteria of this algorithm are the process that requests first are holds the CPU first or which process enter the ready queue first is served first [6]. The workload is processed in the order of arrival time, with no pre-emption [8]. Once a process has been submitted to the CPU, it runs into completion without being interrupted. Such a technique is fair in the case of smaller processes but is quite unfair for long an unimportant job [9]. Since FCFS does not involve context switching therefore it has minimal overhead. It has low throughput since long processes can keep processor occupied for a long time making small processes suffer. As a result waiting time, turnaround time and response time can be low [10]. Shortest Job First (SJF) Scheduling. The criteria of this algorithm are which process having the smallest CPU burst, CPU is assigned to that process next. If two process having the same CPU burst time FCFS is used to break up the tie [6]. SJF can be worked as pre-emptive and non pre-emptive in nature based on the arrival time and burst time of the processes. SJF reduces average waiting time of the processes as compared to FCFS. SJF favors shorter processes over longer ones which is an overhead as compared to FCFS. It selects the job with the smallest burst time ensuing CPU availability for other processes as soon as the current process reaches its completion. This prevents smaller processes from suffering behind larger processes in the ready queue for a longer time [9][11]. Priority Based Scheduling. In this algorithm, priority is associated with each process and on the basis of that priority CPU is allocated to the processes. Higher priority processes are executed first and lower priority processes are executed at the end. If multiple processes having the same priorities are ready to execute, control of CPU is assigned to these processes on the basis of FCFS [1, 4]. Priority Scheduling can be preemptive and non-preemptive in nature. Round Robin (RR) Scheduling. It is a preemptive scheduling algorithm. It is designed especially for time sharing systems. In this algorithm, a small unit of time called time quantum or time slice is assigned to each process [6]. When the time quantum expired, the CPU is switched to another process. Performance of Round Robin totally depends on the size of the time quantum. Longest Wait For Broadcast Scheduling Scheduling. In general algorithms the server process the pages by considering some online policy, and server will process all the outstanding requests separately one after the another. Where as in broadcast scheduling all the requests of the same time are processed / broadcasted at a time. This makes the basic difference between the general algorithms and the broadcast scheduling algorithms. The LWF broadcast scheduling algorithm is motivated from the real situations like wireless and satellite networks, multicast systems and LAN based networks. It is very difficult to develop a algorithm and analyzing a broadcast scheduling even in online or offline for handling simplest method like handling of unit sized pages. In LWF algorithm the author concentrated on minimizing the average flow time, 47

3 waiting time of the requests and other functions. For average flow time three algorithms are known to be O(1)competitive with O(1)speed. The first is the natural longestwaitfirst (LWF) algorithm/policy: at any time t that the server is free, schedule the page p for which the total waiting time of all outstanding requests for p is the largest. Edmonds and Push [16], in a complex and original analysis, showed that LWF is a 6speed O(1)- competitive algorithm and also that it is not O(1)- competitive with a speed less than (1 + 5)/2. The same authors also gave a different algorithm called BEQUIEDF in [17 ] and show that it is a (4 + Є)- speed O(1)competitive algorithm; although the algorithm has intuitive appeal, the proof of its performance relies on an involved reduction to an algorithm for a no clairvoyant scheduling problem [18]whose analysis itself is very complex! The recent improved result in [18] for the no clairvoyant problem when combined with the reduction mentioned above leads to a (2+Є)speed O(1)competitive algorithm; however the new algorithm requires the knowledge of Є and hence is not as natural as the other algorithms. The preemptive algorithms in [17,18] are also applicable when the page sizes are arbitrary; see [19] for empirical evaluation in this model. At a technical level, a main difficulty in online analysis for broadcast scheduling is the fact shown in [20] that no online algorithm can be locally competitive with an adversary. LWF algorithm is concentrated in the setting of unit sized pages. In addition to being a natural greedy policy, it has been shown to outperform other natural policies [2]; moreover, related variants are known to be optimal in certain stochastic settings. It is, therefore, of interest to better understand its performance. The results obtained for the LWF, by conceding all pages of unit sizes are given below LWF is a (4 + Є)speed O(1/Є 2 )- competitive for average flow time via a simple and intuitive analysis. LWF is a (3.4 + Є)speed O(1/Є 3 )- competitive for average flow time via a more complex analysis. 3. Proposed Work ERR Broadcast Scheduling Algorithm Efficient Round Robin algorithm is an online algorithm used in client server technology for broadcasting the pages of a server. In client server technology all the clients are connected to the server, all the clients send the requests to the server for accessing different pages of a server, these pages can be unit sized pages or of different sized pages. The proposed algorithm gives better performance than the existing algorithms. And also it gives better results than the LWF broad cast algorithm. The proposed ERR algorithm can be used in online and offline scenario. It is proven that ERR algorithm [21]see for more details, gives better results in offline scheduling algorithms. LWF algorithm is a broadcast algorithm, and it is shown that this gives better results in broadcast scheduling. The drawback with LWF algorithm is that it is mainly concentrated on processing unit sized pages and it is proffered for online scheduling only. Because broadcast technology always serves multiple clients request whish is known from the real time scenarios like, Wireless networks, LAN and in multicasting process. In LWF the multiple clients request for a page p in a server, the server first process the request, after completion of the process request, and while sending the response to the client, it send the same response to all the clients which request the same page at that time. So all the clients will receive response through that broadcast line. Even the LWF gives the better performance; the drawback is internally the server uses FCFS technique for processing the request. Here our proposed algorithm gives better performance when compared with the LWF algorithm. In our proposed algorithm, the server receives the multiple clients requests, requesting for a page. The server collects the entire request and goes for processing, internally the server uses ERR algorithm. The ERR algorithm uses round robin technique for processing the request. The Round robin algorithm works based on the time slice t. All the process is given equal chance of execution for the time slice amount of time. Round Robin uses cyclic queue so all the request are executed based on the time slice. The request with shortest will be executed first. In ERR algorithm all the processes are placed in the queue in a sorted ascending order. It means all the processes with shortest burst time are executed faster, through which, the algorithm 48

4 increases the execution time, reduces the waiting time and minimizes the flow time. In ERR time slice calculation plays a vital role. Every web server or web container will have some maximum capacity of holding the request, let the maximum capacity of a server or container is m. Suppose if n number of request s come to the server / container, then time slice is calculated as follows. If m > n then time slice t = m/n. Here t value is greater than 1, so all the requests executes with in very less time.. It improves the processing speed of a request. If m < n then time slice t = m / n. Here the time slice is less than 1, so all the requests must wait for some time for completion of execution. Let assume that k-number of clients requests for page number p1, and l-number of clients are requesting for p2 page of a server. Let k1 and l1 are the processes which reaches the server first. Know k1 nad l1 are given for processing using ERR algorithm. It is proven that ERR works better than remaining algorithms [21], so it completes the process request within less amount of time. Now remaining clients process i.e., (k-1) clients requests for page p1, and (l-1) clients requests for page p2 are kept in hold. When the processes k1 and l1 completes their request, and before sending the response to the clients all the requests of the same type i.e. (k-1) of page p1 at that time are broadcasted or transmitted along with the k1 response. Similarly (l-1) requests of page p2 at that time are broadcasted along with the l1 response. So here the container reduces the overhead of processing the similar type of request. Hence ERR algorithm shows better results than that of the LWF broadcast algorithm. ERR scheduling algorithm can work for both online and offline systems depending on the vendors request. 4. Pseudo Code bursttime[n] 0 arrivaltime[n] 0 numprocess[n] 0 turn[n] 0 temp 0 currenttime 0 tslice 0 waittime=0 turntime=0 avgwaititme=0.0 avgturntime=0.0 Read bursttime[n] and arrivaltime[n] For i 0 to n-1 For j I to n If arrivaltime[i] > arrivaltime[j] temp bursttime[i] bursttime[i] bursttime[j] bursttime[j] temp temp arrivaltime[i] arrivaltime[i] arrivaltime[j] arrivaltime[j] temp end for For i 0 to n-1 For j I to n If( arrivaltime == currentime && bursttime[i] >bursttime[j]) temp bursttime[i] bursttime[i] bursttime[j] bursttime[j] temp temp arrivaltime[i] arrivaltime[i] arrivaltime[j] arrivaltime[j] temp end for j wait[n] 0 49

5 if bursttime[i]!= 0 if currenttime == arrivaltime[i] if tslice < bursttime[i] arrivaltime[i] arrivaltime[i]+tslice[i]; given in the simulator. The simulator employs JAVA swings. The front end user interfaces are developed using java awt s and swings. The parent screen allows the user to give number of processes to be in execution. bursttime[i] bursttime[i]-tslice[i] currenttime=currenttime+tslice else arrivaltime[i] arrivaltime[i]+bursttime[i]; currenttime=currenttime+ bursttime[i] bursttime[i] bursttime[i]- bursttime[i] end else Screen2 allows the use give the details of the processes like Burst time, Arrival Time, Priority and Time slice. if bursttime[i] == 0 turn[i] = currenttime end if end if else currenttime++ end else for i 0 to n Screen 3 allows the user to compare any two algorithms of his choice or allows the user to compare all algorithms. wait[i] turn[i]-burst[i]-arrivaltime[i] turn[i] turn[i]-arrivaltime[i] waittime=waittime+wait[i] turnrime=turntime+turn[i] end of for avgwaittime=waittime/n avgturntime = turntime/n Screen 4 shows the comparison results of the selected algorithms. 5. Simulation Design The simulation provides an interactive GUI interface to the user through which a user can input data related to different processes then applying different algorithms based on the algorithm choices 50

6 Comparison of ERR with FCFS Screen 5 gives the average waiting time and average turnaround time of selected algorithms. 6. Compression of ERR with Other CPU Scheduling Algorithms To compare the performance of ERR, it was implemented with some other existing CPU scheduling algorithms. By taking list of processes, processes are scheduled using different CPU scheduling algorithms. Average waiting time and average turnaround time was noted for each. For example the processes are having the following burst time, arrival time and priority. From the above figures it is observed that ERR scheduling algorithm is better than the FCFS scheduling algorithm. Comparison of ERR with SJF Process Burst Time Arrival Time P Priority P P P

7 From the above figures it is observed that ERR scheduling algorithm and the SJF scheduling algorithm are equal. But when time slice is increasing then ERR algorithm gives much better average waiting time and average turnaround time than the SJF algorithm. From the above figures it is observed that ERR scheduling algorithm is better than the ROUND ROBIN scheduling algorithm. Comparison ERR with all Algorithms From the below figures it is observed that ERR Scheduling algorithm works much better than the existing CPU Scheduling algorithms Comparison of ERR with PRIORITY From the above figures it is observed that ERR scheduling algorithm is better than the PRIORITY scheduling algorithm. Comparison of ERR with Round Robin 7. Conclusion The proposed online scheduling algorithm Efficient Round Robin Algorithm can be used as both online and offline scheduling algorithm. ERR algorithm gives better results on a broadcast technology, in 52

8 which multiple clients requests of same type are processed at a single point of time. ERR algorithm reduced the average waiting time, turn around time, flow time of the processes and also increased the performance of a web container by processing the multiple clients requests. It is also proven that ERR gives much better results than existing algorithms like FCFS, SJF, Round Robin, Priority and LWF broadcast algorithm 8. References Abraham Silberschatz, Peter Baer Galvin, Greg Gagne, Operating System Concepts, Sixth Edition. Andrew S. Tanenbaum, Albert S. Woodhill, Operating Systems Design and Implementation, Second Edition. Lingyun Yang Jennifer M. Schopf and Ian Foster, Conservative Scheduling : Using predictive variance to improve scheduling decisions in Dynamic Environments, Super Computing 2003, November 15-21, Pheonix, AZ, USA. Mohammed A.F. Al-Husainy, Best-Job-First CPU Scheduling Algorithm. Information Technology Journal, 6: Sindhu M, Rajkamal R, Vigneshwaran P, An Optimum Multilevel CPU Scheduling Algorithm, ACE, pp.90-94, 2010 IEEE International Conference on Advances in Computer Engineering, P. Balakrishna Prasad Operating Systems second edition. I. A. Dhotre Operating Systems Milan Milenkovic, Operating systems Concepts and Design, McGRAW-HILL, Computer Science Series, second edition. H. M. Dietel, Operating Systems, Pearson Education, Second Edition. M Gary Nutt, Operating Systems A Modern Perspective:, Second Edition, Pearson Education, [14]. Demet Aksoy and Michael J. Franklin. rxw: A scheduling approach for large-scale on-demand data broadcast. IEEE/ACM Trans. Netw., 7(6): , 1999 [15]. Nikhil Bansal, Don Coppersmith, and Maxim Sviridenko. Improved approximation algorithms for broadcast cheduling. In SODA 06: Proceedings of the seventeenth annual ACM-SIAM symposium on Discrete algorithm, pages , [16]. Jeff Edmonds and Kirk Pruhs. A maiden analysis of longest wait first. ACM Trans. Algorithms, 1(1):14 32, [17]. Jeff Edmonds and Kirk Pruhs. Multicast pull scheduling: When fairness is fine. Algorithmica, 36(3): , [18]. Jeff Edmonds and Kirk Pruhs. Scalably scheduling processes with arbitrary speedup curves. In SODA 09: Proceedings of the twentieth annual ACM- SIAM symposium on Discrete algorithm, [19]. Alexander Hall and Hanjo T aubig. Comparing push- and pull-based broadcasting. or: Would microsoft watches profit from a transmitter?. In Proceedings of the 2nd International Workshop on Experimental and Efficient Algorithms (WEA 03), pages , [20]. Bala Kalyanasundaram, Kirk Pruhs, and Mahendran Velauthapillai. Scheduling broadcasts in wireless networks. Journal of Scheduling, 4(6): , [21]. Efficient Round Robin CPU Scheduling Algorithm International Journal of Engineering Research and Development Volume 4, Issue 9 (November 2012), PP J. Wong. Broadcast delivery. Proceedings of the IEEE, 76(12): , S. Acharya, M. Franklin, and S. Zdonik. Disseminationbased data delivery using broadcast disks. Personal Communications, IEEE [see also IEEE Wireless Communications], 2(6):50 60, Dec [13]. S. Acharya, M. Franklin, and S. Zdonik. Dissemination-based data delivery using broadcast disks. Personal Communications, IEEE [see also IEEE Wireless Communications], 2(6):50 60, Dec