In multiprogrammed system, many processes try to read or write the records on disks at the same time. When the requests are made faster, they are serviced either by the moving-head disks, waiting lines or queues build up for each device. In some system, the requests are serviced by simple first-in-first-out basis. First come first served is a fair method of allocating service, but when the request rate becomes heavy, this method leads to long waiting times. If many request arrives, one in the innermost disk, the next to that is in the outermost cylinder, first come first served exhibits a time consuming seeks from the innermost to outermost cylinders.
To avoid time spent in seeking records, a formal and reasonable order of request is needed to queue the request in some manner other than first come first served basis. This process is called disk scheduling.
- FCFS is a basic disk scheduling scheme.
- Through disk scheduling, the requests are keenly observed to find out the efficient way to service the requests.
- A disk scheduler finds the positional relationships between the waiting requests. The request will be serviced in minimum mechanical motion, so such a way the requests need to be reordered.
- There are two types of disk scheduling we are going to discuss. They are:
o Seek optimization
o Rotational or Latency Optimization
- The latency time is the time taken to move the data to rotate from its current position to a position adjacent to the read-write head.
- The seek time is the time taken by the head to move to the cylinder in which the data resides.
- Since seek times tend to be about an order of magnitude greater than latency times, most scheduling algorithms concentrate on minimizing seek times for a set of requests.
- During light-loads, first come first served basis algorithm is sufficient. Only during medium or heavy loading situation, we need to find out some other algorithm with high performance than FCFS.
Characteristics of Disk Scheduling Policies
Mainly, the disk scheduling should have the following three characteristics.
- Throughput : A scheduling policy should have high throughput . Throughput is the number of requests serviced per unit time.
- Mean response time : A scheduling policy should achieve minimum mean response time.
- Variance of response time: variance of response time is one important phenomenon. Variance is a mathematical measure of how far individual items tend to deviate from the average of the items.
- predictability: A scheduling policy should minimize variance. Or else certain requests may experience erratic service levels.
If a scheduling policy merely attempts to improve throughput without minimizing variance simultaneously, it could process the easy-to-service requests and ignore the harder-toservice requests completely.
Types of Disk scheduling
Let us discuss the types of disk scheduling in detail
Seek Optimization – In seek optimization, there are totally six methods. The list of it is given below
- Shortest Seek Time First(SSTF)
- N-Step SCAN
- Eschenbach Scheme.
Want to discuss about the disk scheduling methods in detail? Visit again…