| دانلود رایگان مقاله انگلیسی + خرید ترجمه فارسی
|
|
| عنوان فارسی مقاله: |
الگوریتم زمان بندی جدید برای موازنه (تعادل) بار مزارع سرور |
| عنوان انگلیسی مقاله: |
A new scheduling algorithm for server farms load balancing |
|
|
| مشخصات مقاله انگلیسی (PDF) | |
| سال انتشار | 2010 |
| تعداد صفحات مقاله انگلیسی | 4 صفحه با فرمت pdf |
| رشته های مرتبط با این مقاله | مهندسی کامپیوتر و مهندسی برق |
| گرایش های مرتبط با این مفاله | مهندسی الگوریتم ها و محاسبات، برق مخابرات، برق الکترونیک، نرم افزار و مهندسی فناوری اطلاعات |
| مجله | کنفرانس بین المللی سیستم های اطلاعاتی و صنعتی(International Conference on Industrial and Information Systems) |
| دانشگاه | دانشگاه صنعتی خواجه نصیرالدین طوسی، تهران |
| کلمات کلیدی | مزرعه سرور، توازن و تعادل بار، الگوریتم زمان بندی، اعزام کننده، زمان رفت و برگشت |
| رفرنس | دارد |
| لینک مقاله در سایت مرجع | لینک این مقاله در سایت IEEE |
| نشریه | IEEE |
| مشخصات و وضعیت ترجمه فارسی این مقاله (Word) | |
| تعداد صفحات ترجمه تایپ شده با فرمت ورد با قابلیت ویرایش و فونت 14 B Nazanin | 12 صفحه |
| ترجمه عناوین تصاویر و جداول | ترجمه شده است |
| ترجمه متون داخل تصاویر و جداول | ترجمه نشده است |
| درج تصاویر در فایل ترجمه | درج شده است |
| درج جداول در فایل ترجمه | درج شده است |
- فهرست مطالب:
چکیده
1 مقدمه
2 مزارع سرور وب (وب سرور)
3 الگوریتم های زمان بندی
A FCFS (First-Come-First-Serve)
SRPT B ( کوتاهترین زمان پردازش باقیمانده)
Ps C( به اشتراک گذاشتن پردازش)
D زمان بندی صف چندگانه فیدبک
D نوبت گردشی
E نوبت گردشی وزن دار
4 الگوریتم پیشنهادی
5 شبیه ساز
6 سناریو
7 مقایسه
8 نتیجه گیری
- بخشی از ترجمه:
الگوریتم کوتاهترین زمان پردازش باقیمانده (SRPT) در میان الگوریتم های زمان بندی که از گرسنگی رنج می برند، بهترین عملکرد را به معرض نمایش می گذارد. الگوریتم FCFS منصفانه است اما متوسط زمان رفت و برگشت، بسیار بالا است. هر الگوریتم زمان بندی که شباهت بیشتری به الگوریتم SRPT دارد، عملکرد بهتری نشان می دهد. الگوریتم پیشنهادی تشریح شده در این مقاله از الگوریتم HRRN در اعزام کننده مزارع سرور برای توزیع کارهای ورودی میان سرورها استفاده می کند. حسن استفاده از این الگوریتم، عملکرد بهتر می باشد. متوسط زمان رفت و برگشت الگوریتم HRRN شباهت بیشتری به SRPT دارد. اصلی ترین حسن الگوریتم HRRN غلبه بر گرسنگی می باشد. این الگوریتم از گرسنگی رنج نبرده و دارای عملکردی یکسان با SRPT می باشد. بنابراین استفاده از الگوریتم HRRN در اعزام کننده مزارع سرور می تواند عملکرد را افزایش دهد.
- بخشی از مقاله انگلیسی:
Long round around time is a very important problem in internet services. There is more and more evidence showing a high variability of task size distribution in computer loads. For example, files requested in Web servers and in UNIX fit a heavy-tailed distribution [5]. One way to figure out this problem is using server farms. Server farm is consisting of a collection of many computers also called host, server or node and front-end high-speed dispatcher. Each incoming request is immediately dispatched via the dispatcher to one of the computers. The main advantages of using server farms are price and flexibility, because many slow computers are cheaper than fast computers and it is easy to up or down your server capacity. One of the most important issues in server farms is “routing policy”, also known as “task assignment policy”. This is the algorithm/rule for determining how to assign jobs to hosts. On the Internet, companies whose web sites get a great deal of traffic usually use load balancing. For load balancing Web traffic, there are several approaches. For Web serving, one approach is to route each request in tum to a different server. In some approaches, the servers are distributed over different geographic locations. For distributing incoming jobs on the several servers, dispatcher is used. Dispatcher objective distributes incoming job on the servers in an efficient way. Dispatcher use scheduling algorithms to be able to dispatch jobs in the best way. Therefore scheduling algorithms is very important to achieve better performance in dispatcher. This paper describes a novel and efficient scheduling algorithm to be used in dispatcher for better job distributing in order to increase server farms performance. 978-1-4244-8217-7110/$26.00 ©201O IEEE Shafigh Parsazad Ferdowsi U ni versity of Mashhad Mashhad, Iran parsazad @stu.um.ac.ir II. WEB SERVER FARMS A Web server farm refers to a Web site that uses two or more servers housed together in a single location to handle user requests. They use one host called site to provide a single interface for users [1]. Today busy web servers are required to service many clients simultaneously, sometimes up to tens of thousands of concurrent clients [8]. Large web-server farms consist of thousands of servers and may handle millions of HTTP requests per second. These sites are overwhelmed by the offered load and the Web service provider shave to deal with peak demands that are much higher than the average load supported by their site. While operators are just about to be able to collect detailed traffic statistics, the very detail and volume make them nearly impossible to analyse. As a result, performance prediction, monitoring and performance measurement are rendered increasingly complicated, one computer cannot be able to handle the requests, no matter how many disks are used in parallel. The slow response times and difficult navigation are the most common complaints of Internet users [9]. Research shows the need for fast response time. The response time should be around 8 seconds as the limit of people’s ability to keep their attention focus while waiting [10]. To figure out this problem more Computers must be added. Server farms consist of two important parts. First part is called “Front-End “and second is called “Back-End”. To deliver requests to the individual machines of a server farm, a device is needed to accept all incoming traffic and to assign arequest to a particular machine to handle [3]. These devices are called Front End Devices (FEDs), because they sit at the front of a server farm accepting all requests. A major problem now arises, if a similar workstation or PC is used for a FED, then the number of attached machines in the server farm is typical limited to ten or less [4]. The most important part of any Web farms is the part that assigns the incoming load. This entity is called “Dispatcher”. In other words, the Dispatcher acts as a centralized global scheduler that receives the totality of the requests and routes them among the back-end servers of the Web farm [2]. The dispatcher may use different scheduling policies to assign the load to the nodes of Web servers. Each server machine in the web farm is uniquely identified with a private address to access. Figure 1 is illustrated the server farm model. The problem with web server farms is that a load imbalance may occur, where some processors are idle with nothing to execute, while other processors are busy and have tasks in their ready queues. Any load imbalance will result in poorer task response times. Scheduling policies determine which requests in the queue are serviced at any point of time, how much time is spent on each, and what happened when a new request arrives. The goals of scheduling policies are to minimize the mean round around time of the request and to behave fairly to all requests [2]. Static algorithms are the fastest solutions. They do not have overloading decision making time. Static algorithms can potentially make poor assignment decisions, such as routing a request to a server node havinga long queue of waiting load while there are other almost idle nodes [5]. Dynamic algorithms have the potential to outperform static algorithms by using some state information to help dispatching decisions. On the other hand, they require mechanisms that collect, transmit and analyse state information there by incurring in overheads [1]. Round-Robin (RR), random (RAN) and Weighted Round-Robin (WRR) are dynamic algorithms.
| دانلود رایگان مقاله انگلیسی + خرید ترجمه فارسی
|
|
| عنوان فارسی مقاله: |
الگوریتم زمان بندی جدید برای موازنه (تعادل) بار مزارع سرور |
| عنوان انگلیسی مقاله: |
A new scheduling algorithm for server farms load balancing |
|
|
