Synopsis :- Distributed Process Scheduling

1995 days ago, 675 views
PowerPoint PPT Presentation
Rundown of section 5. A System execution modelStatic process schedulingDynamic burden sharing and balancingDistributed process implementationReal time planning. A framework execution model[2]. Portrays the relationship among calculation, planning and structural engineering to depict the Inter process communicationBasically three sorts of model are there:Precedence process model(DAG) Directed edges re

Presentation Transcript

Slide 1

Outline :- Distributed Process Scheduling Prepared BY:- JAYA KALIDINDI

Slide 2

Summary of part 5 A System execution display Static process planning Dynamic load sharing and adjusting Distributed process usage Real time booking

Slide 3

A framework execution model[2] Depicts the relationship among calculation, booking and engineering to portray the Inter procedure correspondence Basically three sorts of model are there: Precedence handle model(DAG) Directed edges speak to the priority relationship

Slide 4

Communication prepare show In this model procedures exist together and Communicate synchronously. Edges in this model speak to the need of correspondence between the procedures

Slide 5

Disjoint process display: In this procedures run freely and finished in limited time. Procedures are mapped to the processors to augment the usage of procedures and limit the turnaround time of the procedures.

Slide 6

Speed up N-Number of procedures -Efficiency Loss when executed on a genuine machine. RC-relative simultaneousness RP-relative preparing necessity Speed up relies on upon: Design and effectiveness of the booking calculation. Engineering of the framework

Slide 7

Static Process Scheduling Is utilized to discover ideal answer for the issue. There are two extraordinary instances of work task. mapping of procedures is done before execution of the procedures. when process began it remains at the processor until consummation of the procedure. What's more, never acquired. Decentralized and non –Adaptive are the disadvantages of Static process booking.

Slide 8

Dynamic load sharing and Balancing Load adjusting can be characterized as a procedure to disperse work between numerous PCs forms or some other assets to get ideal asset use. controller diminishes the procedure lingering through load sharing ie,by joining the most limited line and evening out line sizes by load adjusting . Advance, procedures can be permitted to move from longer line to shorter line through load Redistribution.

Slide 9

Sender Initiated Algorithm It is actuated by a sender procedure that desires to off-load some of its calculation by relocation of procedures from a vigorously stacked sender to a gently stacked recipient. Exchange of process shape a sender to reciever requires three essential decisions. Transfer arrangement:- when does the hub turn into the sender? Choice Policy:- How does the sender pick a procedure for exchange? Area Policy:- which hub ought to be the objective reciever?

Slide 10

Receiver started Algorithm:- This are the draw models in which beneficiary can pull a procedure from others to its site for execution. They are more steady than the sender started calculation. At high framework stack ,handle movements are few and a sender can be discovered effortlessly. Recipient started calculations perform superior to anything the sender started calculations Both the calculations can be joined relying upon RT and ST.

Slide 11

Distributed process execution Remote Service :- The message is translated as a demand for a known administration at the remote site Three diverse programming levels:- As remote strategy calls at the dialect level. As remote charges at the working framework level. As interpretive messages at the application level.

Slide 12

Distributed process Implementation Depending on how the demand messages are deciphered, there are three primary application situations: Remote Service The message is translated as a demand for a known administration at the remote site. Remote Execution The messages contain a program to be executed at the remote site. Prepare Migration The messages speak to a procedure being relocated to a remote site for proceeding with the execution.

Slide 13

Remote Execution The reason for remote administration is to get to the remote host not at all like remote administration remote process keeps up the perspective of beginning framework. Some Implementation issues:- stack sharing calculations. Area autonomy. Framework heterogeneity. Insurance and security.

Slide 14

Load-Sharing Algorithm Each procedure server are dependable to keep up the heap data. The rundown of hosts taking an interest are communicated. The determination system is by a unified merchant handle. Once a remote host is chosen The customer procedure server demonstrates the asset necessities to the procedure server at the remote site. On the off chance that the customer is confirmed and its asset necessities can be met, the server gifts consent for remote execution. The exchange of code picture takes after, and the server makes the remote procedure and the stub. The customer introduces the procedure forked at the remote site.

Slide 15

Location Independence Process made by remote execution obliges coordination to finish basic undertaking. So it is important to bolster sensible perspectives for the procedures. Every remote procedure is spoken to by a specialist procedure at the beginning host. It seems like the procedure is running on a solitary machine.

Slide 16

System heterogeneity If remote execution is summoned on heterogeneous host , then it is important to re-arrange the program. Overhead Issue. Arrangement: Use accepted machine-free transitional dialect for program execution.

Slide 17

Process Migration The message speaks to a procedure being relocated to the remote site for proceeding with execution. Prepare movement office State and setting exchange:- It exchanges the calculation state data and correspondence state data

Slide 18

Real Time Scheduling:- The frameworks which safeguards that specific moves are made inside indicated time limitations are called constant frameworks. Can be delegated: Static versus dynamic Premptive versus non-premptive Global versus Local

Slide 19

Rate Monotonic It's anything but difficult to actualize. Sorts the assignments by the lengths of their periods. It likewise makes great need assignments. Rate monotonic is an ideal need task calculation.

Slide 20

Deadline Monotonic :- progressively framework a few undertakings need to finish execution a brief span in the wake of being asked. Soonest Deadline First :- this is applies dynamic need booking to accomplish better CPU use . Constant Synchronization:- An arrangement of undertakings that participate to accomplish an objective should share data and assets or different words synchronize with different assignments.

Slide 21

References:- [1].[2]. Randy Chow, Theodore Johnson, "Dispersed Operating Systems & Algorithms", Addison Wesley.(all outlines) [3]. Dejan S. Milojicic Fred Douglis Yves Paindaveine Richard Wheeler Songnian Zhou, "Handle Migration" , ACM Computing Surveys (CSUR) Volume 32 ,  Issue 3  (September 2000) [4]. S. Cheng, J.A. Stankovic and K. Ramamritham, ''Scheduling Algorithms for Hard Real-Time Systems: A Brief Survey'', page6-7 in Hard Real-Time Systems: Tutorial , IEEE (1988). [5] .Distributed Process Scheduling . Progressed Operating Systems Louisiana State University Rajgopal Kannan. Issues in Distributed Scheduling .

Slide 23