Have you ever wondered how multiple I/O requests made simultaneously to the Operating System are serviced? In a disk controller, only one I/O request can be serviced at a time. Suppose we have two important I/O requests from different processes seeking to use the disk at a particular instance. How does the disk controller decide the request that gets serviced and the request that stays in the ready queue? There are several ways to determine how the disk attends to multiple I/O requests; they are known as disk scheduling algorithms.
In this blog post, you will learn about disk scheduling algorithms, the different disk scheduling algorithms, and the pros and cons of each type.
What Are Disk Scheduling Algorithms?
Disk Scheduling Algorithms refer to the particular manner in which the operating system schedules I/O requests arriving to be serviced by the disk.
It is important to define these scheduling algorithms to avoid having a clustered hard drive and to efficiently access data. For instance, having a defined service sequence for requests defeats the need of unnecessarily waiting in the queue. The major purpose of implementing disk scheduling algorithms is to reduce the seek time.
Seek Time refers to the time taken for the disk controller to locate a piece of stored data. The seek time can vary based on the type of scheduling algorithm used and the distance from the start point to where the read/write head of the disk is. Now let's compare disk scheduling algorithms.
Types of Disk Scheduling Algorithms
In this section, we will discuss the 6 major types of disk scheduling algorithms, their advantages, as well as their disadvantages.
NB: Throughout the course of this post, we set our read/write head at track 50 and our tail track at track 199.
First Come First Serve (FCFS)
As the name implies, the First Come First Serve scheduling algorithm services requests in order of their arrival. That is, the request that arrives first is serviced first, and then subsequent requests are also placed in a queue in order of their arrival. Consider the FCFS disk scheduling algorithm example below:
Here, the requests arrive in this order: (100,180,40,150,80,15,190).
To find the total head movements, we calculate the number of tracks it took from one request to another.
Hence, the total head movements for the FCFS algorithm above is given by = (100-50)+(180-100)+(180-40)+(150-40)+(150-80)+(80-15)+(190-15) = 690.
Advantages of FCFS
- It is the simplest disk scheduling algorithm and easy to implement.
- Since all requests are serviced in order of their arrival, there is no starvation of any kind. Starvation is a problem where a process does not get executed due to a lack of the needed resources. It occurs when high-priority processes continually get executed thereby causing low-priority processes to wait indefinitely.
- All requests are treated fairly and eventually get attended to.
Disadvantages of FCFS
- Seek time is excessively increased using the FCFS disk scheduling algorithm.
- It falls short in situations where some requests have to be processed with high priority since FCFS does not provide any way to distinguish between these requests.
- It is very inefficient especially when dealing with a high number of requests.
Shortest Seek Time First (SSTF)
The Shortest Seek Time First disk scheduling algorithm processes requests based on the length of their seek time. That is, the request with the shortest seek time gets serviced first. In the execution of the SSTF algorithm, the disk controller performs a quick scan of all the requests to be serviced and arranges them for processing in order of their seek time. The SSTF disk scheduling algorithm addresses the limitations of the FCFS by prioritizing the reduction in total seek time. The disk controller then services the request with the shortest seek time from its current position.
Suppose the requests arrive in this order: (100,180,40,150,90,15,190). Here's how the disk controller processes these requests:
In the sample diagram above, we see that from the disk head (50), the shortest next distance from the disk head is 40. Note that the request in track 100 arrives first but is not executed first; the request in track 40 is only 10 tracks away from the disk head, which is the shortest and ultimately how the SSTF algorithm works. From track 40, it moves to track 15, which is the shortest next distance being only 25 tracks away and then it progresses that way till it gets to the last track.
The total head movement is given by = (50-40)+(40-15)+(90-15)+(100-90)+(150-100)+(180-150)+(190-180)=210.
We see that this is a significantly reduced average seek time in comparison to the FCFS algorithms.
Advantages of SSTF
- It has a significantly reduced average seek time.
- It provides increased throughput. Throughput refers to the number of processes completed per unit time.
Disadvantages of SSTF
- One major setback of using the SSTF algorithm is the risk of starvation of a request with a higher seek time compared to other requests.
- There is a high variance in response time and waiting time. Response time refers to the amount of time it takes the disk to respond to a particular request while waiting time refers to the total amount of time a request spends in the ready queue before reaching the disk.
- There is an overhead in finding the closest request since it has to calculate the seek time of all requests in advance.
This is also known as the elevator algorithm because of its working principle. The SCAN disk scheduling algorithm scans down in a single direction to the nearest end from the disk head and then reverses its direction and services all the other requests in the opposite direction.
Due to the working principle of the SCAN algorithm, requests placed at the midrange position from the disk head have a higher chance of getting serviced than those at the end of the disk head and definitely a higher chance than requests placed behind the disk head. Let's consider an example where our requests arrive in this order: (100,180,40,150,90,15,190). Using the SCAN algorithm in disk scheduling, the requests are serviced in this manner:
Here, we see the requests to the left end are initially serviced, and then it reverses its direction (as we discussed) to service requests on track 40 and track 15.
The total overhead movement is given by: (90-50)+(100-90)+(150-100)+(180-150)+(190-180)+(199-190)+(199-40)+(40-15)=333. Alternatively, we can calculate the overhead movement by subtracting the tracks of the disk head(50) and final point(15) from the tail track(199) where it reverses its direction. We have: (199-50)+(199-15)=333
Advantages of SCAN Algorithm
- It is easy to understand.
- The SCAN algorithm provides a high throughput.
- It has a low variance in response time and waiting time.
- There is a very low risk of starvation.
Disadvantages of SCAN Algorithm
- Requests just visited by the disk head will have a relatively longer waiting time.
- Sometimes, it may not be optimal because it causes the disk head to move to the end of the disk by default, whether or not there are requests to be serviced.
Circular Scan (C-SCAN) Algorithm
The C-SCAN algorithm addresses the limitations of the SCAN algorithm. Unlike the SCAN algorithm that rescans requests that have been serviced on the first run, the C SCAN disk scheduling algorithm returns to the starting point without scanning (again) any requests in between and then begins to service requests from the other end. Suppose we have our requests in this manner: (100,180,40,150,90,15,190). Using the C-SCAN algorithm, here's how the requests are serviced:
Here, we see that when the requests are serviced to the end point, it behaves similarly to the SCAN algorithm, but when it reverses its direction to the start point, it does not service any request until it gets to the start point; when it gets to the start point, it then services the request in the same direction.
The total overhead movement is given by:(90-50)+(100-90)+(150-100)+(180-150)+(190-180)+(199-190)+(199-0)+(15-0)+(40-15)= 388. Or simply: (199-50)+(199-0)+(40-0)=388
Advantages of C-SCAN
- It provides a more uniform and accurate waiting time since all requests will be serviced once.
- It provides a better waiting time than the SCAN algorithm.
Disadvantages of C-SCAN
- Like the SCAN algorithm, it causes the disk head to move to the end of the disk even if there are no available requests to be serviced.
The Look Disk Scheduling Algorithm is similar to the SCAN algorithm. In the LOOK algorithm, the disk head moves to the last request (and not the end of the disk as is the case for the SCAN algorithm) to be serviced in front of the head, then it reverses its direction and services all the requests behind the disk head. This algorithm addresses the limitation of the SCAN algorithm by preventing the extra delay used in traversing to the end of the disk when there are no requests to be serviced. Let's see how the requests: (100,180,40,150,90,15,190) will be serviced using the LOOK algorithm:
Here, we see that it traverses to track 190 (which is the furthest request to be serviced) and not track 199 (the end of the disk track). Then it reverses its direction and services the requests on track 40 and track 15.
The total overhead movement is simply given by: (190-50)+(190-15)=315
Advantages of LOOK Algorithm
- It saves time that may have been wasted in unnecessarily traversing to the end of the disk when there are no requests.
- It is efficient because only locations with data in it are visited.
- It provides low variance in response time and waiting time.
Disadvantages of LOOK Algorithm
- There is an overhead in locating the end requests.
- Requests that are located far away from the disk head are prone to starvation.
Circular Look (C-LOOK) Algorithm
The C-LOOK algorithm is similar to the C-SCAN algorithm as it does not rescan requests that have already been visited on the first run to the end of the disk. The C LOOK disk scheduling algorithm allows the disk arm to reverse its direction and traverse to the first request (not the start point like the C-SCAN algorithm). Here's how a similar order of requests will be serviced using the C-LOOK algorithm:
Here, just like the C-SCAN algorithm, it does not service any request (again) when it reverses its direction. The remaining requests are serviced after the disk arm reaches the start of the leftmost request.
The total overhead movement is given by: (190-50)+(190-15)+(40-15)=340
Advantages of C-LOOK Algorithm
- Compared to LOOK algorithm, it provides a better performance since it does not service requests that have already been serviced.
- Starvation of requests is avoided.
- It provides low variance in response time and waiting time.
Disadvantages of C-LOOK Algorithm
- There is an overhead in finding end requests.
Disk Scheduling Algorithms x Pieces Copilot — Getting a Head Start
Perhaps you may not fancy drawing a graph or manually computing request parameters to get the seek time and determine the order the requests get serviced, you can simply get these by asking the Pieces Copilot. In the example below, we provide an order of requests to the Pieces Copilot and ask it how the requests will be serviced by a disk controller using the Shortest Seek Time First Algorithm. Here's the conversation:
And Voila! It shows us the order just like we represented in our diagram we explained in the SSTF section. Ready to enhance your productivity and get ahead of it all using the Pieces Copilot? Then download the Pieces for Developers Desktop App here and get started!
By following this guide, you have learned about disk scheduling algorithms, the various types of disk scheduling algorithms, a comparison of disk scheduling algorithms, as well as how to get ahead and enhance your productivity using the Pieces for Developers Desktop App. Have fun!