Computer System Architecture
Summary
TLDRThis lecture delves into computer system architecture, focusing on categorizing systems by the number of general-purpose processors. It introduces single processor systems with a main CPU and optional special purpose processors for device-specific tasks. Multiprocessor systems, offering increased throughput and reliability, are explained, highlighting symmetric and asymmetric multiprocessing. Lastly, clustered systems, composed of multiple interconnected systems for high availability, are discussed, contrasting symmetric and asymmetric clustering approaches.
Takeaways
- đ The lecture focuses on computer system architecture, specifically categorizing computer systems based on the number of general-purpose processors.
- đ» Single processor systems have one main CPU capable of executing a general-purpose instruction set, including user process instructions, and may also include special-purpose processors for device-specific tasks.
- đ Multiprocessor systems contain two or more processors that are closely communicating and sharing resources like the computer bus, memory, and peripheral devices.
- đ Multiprocessor systems offer advantages such as increased throughput, allowing for more data to be processed in parallel, and thus improving system performance.
- đ° The economy of scale is a benefit of multiprocessor systems, as multiple processors can share resources, making them more cost-effective than multiple single-processor systems with separate resources.
- đĄïž Multiprocessor systems are more reliable than single-processor systems because if one processor fails, the others can continue to operate, preventing a total system failure.
- đ There are two types of multiprocessor systems: symmetric multiprocessing, where all processors are the same and share tasks, and asymmetric multiprocessing, which has a master-slave approach with a designated master processor.
- đ Clustered systems are composed of two or more individual systems coupled together, providing high availability by allowing other systems to take over tasks if one fails.
- đ Like multiprocessor systems, clustered systems can be structured either symmetrically or asymmetrically, with symmetric structures allowing for more efficient resource sharing.
- đ In asymmetric clustered systems, one machine operates in hot-standby mode, ready to take over if another system fails, while in symmetric systems, all systems monitor each other and share tasks.
- đ The choice between symmetric and asymmetric structures often leans towards symmetric for better resource utilization and efficiency.
Q & A
What are the main types of computer systems based on the number of general purpose processors?
-The main types of computer systems based on the number of general purpose processors are single processor systems, multiprocessor systems, and clustered systems.
What is a single processor system?
-A single processor system is a computer system that contains one main CPU capable of executing a general purpose instruction set, including instructions from user processes, along with special purpose processors for device-specific tasks.
What is the difference between a general purpose processor and a special purpose processor?
-A general purpose processor is capable of executing a wide range of instructions and handling various tasks, while a special purpose processor is designed to perform specific tasks related to a particular device or function.
What are the advantages of a multiprocessor system over a single processor system?
-The advantages of a multiprocessor system include increased throughput, economy of scale, and increased reliability due to the presence of multiple processors that can share workload and provide backup in case of failure.
What is the meaning of 'throughput' in the context of computer systems?
-In the context of computer systems, 'throughput' refers to the measure of the performance of the system, often defined as the amount of data that can be transferred from one location to another.
What is the concept of 'economy of scale' in relation to multiprocessor systems?
-'Economy of scale' in relation to multiprocessor systems refers to the cost efficiency gained from sharing resources among multiple processors, as opposed to having separate single processor systems with individual resources.
How does a multiprocessor system provide increased reliability?
-A multiprocessor system provides increased reliability because if one processor fails, the remaining processors can continue to operate and share the workload, preventing a total system failure.
What are the two types of multiprocessing systems?
-The two types of multiprocessing systems are symmetric multiprocessing (SMP) and asymmetric multiprocessing (AMP).
What is the key difference between symmetric and asymmetric multiprocessing?
-In symmetric multiprocessing, all processors are similar and participate equally in task execution, whereas in asymmetric multiprocessing, there is a master-slave approach where one processor acts as a master and the others as slaves with specific tasks assigned by the master.
What is a clustered system, and how does it differ from a multiprocessor system?
-A clustered system is a computer system that couples two or more individual systems together to accomplish computational work. Unlike a multiprocessor system, which consists of multiple processors, a cluster system involves complete systems working in unison, providing high availability and redundancy.
How does a clustered system provide high availability?
-A clustered system provides high availability because multiple systems are working together; if one system fails, the others can take over the tasks, minimizing the chance of a complete system failure.
What are the symmetric and asymmetric structures in a clustered system?
-In a symmetric clustered system, all hosts run applications and monitor each other, sharing resources efficiently. In an asymmetric clustered system, one machine operates in hot-standby mode while the others run applications, with the standby machine ready to take over if another fails.
Outlines
đ» Introduction to Computer System Architecture
This paragraph introduces the topic of computer system architecture, focusing on the categorization of computer systems based on the number of general-purpose processors. It explains the three main types: single processor systems, which contain one main CPU for executing general-purpose instructions and may include special-purpose processors for device-specific tasks; multiprocessor systems, featuring two or more processors that work closely together and share resources; and clustered systems, where multiple systems are grouped to perform tasks more efficiently. The paragraph sets the stage for a deeper exploration of these systems and their advantages.
đ Advantages of Multiprocessor Systems
The second paragraph delves into the advantages of multiprocessor systems over single processor systems. It highlights increased throughput, which is the system's ability to handle a larger amount of data transfer, resulting in faster and more efficient performance due to parallel processing. The economy of scale is also discussed, emphasizing the cost-effectiveness of multiprocessor systems as resources are shared among multiple processors, unlike single processor systems that require individual resources. Additionally, multiprocessor systems offer increased reliability, as the failure of one processor does not lead to a total system failure, with other processors able to take over the workload and maintain system operation, albeit with reduced performance.
đ Types of Multiprocessor Systems
This paragraph explores the two types of multiprocessor systems: symmetric multiprocessing (SMP) and asymmetric multiprocessing (AMP). In SMP, all processors are identical and participate equally in task execution, while in AMP, there is a master-slave relationship where one processor acts as a master, overseeing and assigning tasks to the slave processors. The paragraph explains the functioning of each type and the roles of the processors within them, illustrating the differences in their structures and operation.
đ Understanding Clustered Systems
The final paragraph discusses clustered systems, which are composed of multiple individual systems working together to perform computational tasks. Unlike multiprocessor systems, which consist of multiple processors, clustered systems involve complete systems that are interconnected. The paragraph explains the concept of high availability in clustered systems, as the failure of one system does not result in total system failure due to the redundancy provided by other systems in the cluster. It also touches on the symmetric and asymmetric structures of clustered systems, with symmetric offering more efficient resource sharing and utilization compared to the master-slave approach of asymmetric clustering.
Mindmap
Keywords
đĄOperating System
đĄGeneral Purpose Processor
đĄSingle Processor System
đĄMultiprocessor System
đĄClustered System
đĄThroughput
đĄEconomy of Scale
đĄReliability
đĄSymmetric Multiprocessing
đĄAsymmetric Multiprocessing
đĄHigh Availability
Highlights
Introduction to computer system architecture based on the number of general purpose processors.
Definition and explanation of single processor systems with one main CPU for general purpose tasks.
Identification of special purpose processors for device-specific tasks in single processor systems.
Clarification that the term 'single processor system' refers to the number of general purpose processors, not the total number of processors.
Introduction to multiprocessor systems with two or more processors for parallel processing.
Description of multiprocessor systems as parallel or tightly coupled systems sharing resources.
Advantage of increased throughput in multiprocessor systems due to parallel processing.
Economic benefits of multiprocessor systems through shared resources compared to single processor systems.
Increased reliability in multiprocessor systems as a result of having multiple processors to compensate for failures.
Differentiation between symmetric and asymmetric multiprocessing systems.
Symmetric multiprocessing characterized by similar processors sharing tasks equally.
Asymmetric multiprocessing featuring a master-slave approach with distinct roles for each processor.
Introduction to clustered systems as a collection of individual systems working together.
High availability in clustered systems due to multiple systems compensating for failures.
Structuring of clustered systems as either asymmetric with a hot-standby mode or symmetric with mutual monitoring.
Comparison of efficiency in symmetric vs. asymmetric clustered systems, favoring the former for better resource sharing.
Summary of the types of computer systems based on the number of general purpose processors and their advantages.
Transcripts
In the previous lectures,
we have studied the basic of the Operating system.
And we have seen things like computer system operation,
and we have also seen structures like storage structures
and input/output structures.
Now, in this lecture, we will study the computer system architecture
and mainly about the types of computer systems
based on number of general purpose processors.
That means we are going to study how we can categorize the computer system
based on the number of general purpose processor.
Alright!
So, let's see how we categorize the computer system
based on the general purpose processor.
So, first of all, we have a single processor system.
Which from the name itself you may have understood
that it has only a single processor.
And then we have the multiprocessor system
in which there are two or more processors.
And then we have the clustered system.
In which two or more systems are coupled
or clustered together in order to perform certain task.
Now, let us explain each one of these systems
and see what are the differences and what are the advantages of them,
and what are the types of these processing systems that we have.
So, first of all, we have the single processor system.
So, in the single processor system one main CPU
capable of executing a general purpose instruction set
including instructions from user processes.
That means in the single processing system we only have one main CPU,
which is capable of executing the general purpose instruction set
as well as the instructions from the user processes.
Now, other special purpose processors are also present
which perform device specific tasks.
Now, in the single processing system, apart from the main CPU
there are also other processors which are present
which don't do the general purpose task but perform some device specific task.
Now, what do we mean by this?
This means that we have certain devices in our computer,
like keyboard, hard disk, and so on.
So for all this, there may be some microprocessor
that is specified to do a specific task related to that device.
Like for example, let's say that we have our keyboard.
Now when you pressed a key on your keyboard,
the key stroke has to be converted to some kind of code.
let's say a binary code.
so that the computer actually understands
what you have pressed or what you are typing.
So, in order to convert a keystroke to a code
there is a little microprocessor present on your keyboard.
which is going to perform only that task of
converting your keystroke to some kind of code.
Now, those are the special purpose processors that we are talking about.
So, these kind of special purpose processors are also present
but they are not for general purposes but specific purposes.
Now, You may question me saying that
I said it is a single processor system
but here I have already shown you more than one processor.
I said that there are general purpose processors
and also special purpose processors.
yes, that is right!
But the answer to this is that
we give the name to this processor system
based on the number of general purpose processors that it has.
So even though there are special purpose processors
which perform some little task specific to the devices.
But as we have only one general purpose processor
we called it a single processor system.
So, that is why we called it the single processor system.
Alright! now let's go to the next one.
So, the next one is the multiprocessor system.
So, from the name itself, we can understand that
in this there will be more than one processor.
Or we can say that there are two or more processors.
And also it is known as a parallel system or a tightly coupled system.
So this multiprocessor system is also known as
a parallel system or a tightly coupled system.
and these two terms will be discussing later as we move ahead in the subject.
So, in these multiprocessor systems
they have two or more processors in close communication
sharing the computer bus
and sometimes the clock, memory, and peripheral devices.
So, unlike the single processor system which has only a single processor
In this multiprocessor system, we have two or more processors,
which are closely communicating with each other
because they need to work together to perform certain tasks.
And what are those tasks?
There may be a single task or different tasks
depending on the kind of system that we have.
So in order to work together,
they need to have close communication and synchronization with each other.
So, that is why we say that they are in a close communication
and they also share the computer resources that we have.
Now, let us see what are the advantages of having this multiprocessor system
or single processor system.
So, coming to the advantages.
The first advantage is to increased throughput.
So, throughput is something that we can use
to measure the performance of our system.
So, sometimes it is also known as the amount of data
that can be transferred from one location to another.
So, throughput is like the measure of the performance of the system.
So, we can say that in the multiprocessor system,
we have more throughput or more performance.
And why is that?
The answer is simple, that is because we have more than one processor
which will parallelly do our work
and it will make it faster and more efficient.
And that is what we mean by increased throughput.
And then the second advantage is economy of scale.
Now, why do we call it economic?
I say that multiprocessor systems are more economical
as compared to the single processor system.
That is because in the multiprocessor system
we have these different processors
sharing the resources of our computers or our systems.
But in the case of a single processor system,
if you want to have the same amount of power
you may need multiple single processor system
with their own individual resources.
But here we see that we have multiple processors sharing the resources
So, in that way, we can have economy of scale.
That means it is more economic as compared to single processor system.
So, if you have multiprocessor system with three processors
and if you want to have same kind of set up in your single processor system
You may need to have three single processor systems.
three individual single processor system,
with their own individual resources.
We see that this is more economic.
then the other advantage is increased reliability.
this is more reliable as compared to the single processor system.
Why is that?
The reason is again the same that is because we have more than one processors.
And why it is reliable?
that is because in these multiprocessor system, even if a single processor fails,
we still have other processors which will back us up,
and which will still keep us running.
So, let's say that you have ten processors in your ten multiprocessor system
and let's say that one of them fails.
So, even if one of them fails,
the remaining nine can share the work that was being done by the one that is failed
and without giving you complete failure,
your system will still work even though the performance may reduces a bit.
But in case of single processor system,
If your single processor fails,
that means it is a total failure, your whole system breaks down.
But in multiprocessor system, even if one fails
the remaining can help to get the work done
without causing a total failure of the system.
So, that is why we say that it is more reliable.
It has increased reliability
as compared to the single processor system.
Alright!
So, those were the advantages of the multiprocessor system.
Now, let's see what are the types of the multiprocessor systems that we have.
So, basically there are two types of multiprocessor systems.
first one is called symmetric multiprocessing.
and the second one is asymmetric multiprocessing.
Now, let us see what are the difference between these two
types of multiprocessing systems.
So, in symmetric multiprocessing,
We have our different processors.
here CPU one, two, three denote the different processes that we have.
And processes P1, P2, P3 denote the processes that have to be executed.
Now, in symmetric multiprocessing
all these CPUs and processors are actually the same.
that means they are similar to each other.
and they all participate in performing these task or processes P1, P2, and P3.
So, all the CPUs are involved in these tasks.
So, that is known as symmetric multiprocessing.
Now, in asymmetric multiprocessing,
unlike these one where CPU one, two, and three are similar to each other.
Here we have a master slave approach.
So here these boxes denote the CPUs.
So, one of the CPU and one of the processor will act as a master,
and the remaining processors will be slaves.
So, the master monitors the other processors.
and it assigns the task for these processors.
So, the master is like the monitor which is
guiding and supervising the other slaves.
and then the slaves take care of particular processes.
so, let's say, here slave1 is taking care of process1,
slave2 for P2,
and slave3 for P3.
and we see that slave1 only taking care of the process P1.
and process P1 is not taking care by slave2 or slave3.
So, P1 is only for slave1.
and so is P2 and P3 for slave2 and slave3 respectively.
and here also if something fails,
the master will take care of how to distribute the loads
after one fails and things like that.
So, the master is the monitor,
and the rest are the slaves.
But in symmetric multiprocessing,
we see that there is no master slave approach,
but all the processors are peers or they are similar to each other.
and they are all involved in all the task that is present.
so, that is the difference between
symmetric multiprocessing and asymmetric multiprocessing.
Now, let us come to the next type of computer system
that we have based on the number of processor
which is the clustered system.
So, in this clustered system,
like multiprocessor systems, clustered system gathered together
multiple CPUs to accomplish computational work.
So, this is also somewhere similar to our multiprocessor system.
So, here the cluster system, they gather together
multiple CPUs to accomplish computational work.
Now you may be thinking if that is the case,
then what is the difference between
multiprocessor system and clustered system?
So, the difference is that
they are composed of two or more individual systems coupled together.
So, in multiprocessor system what we had was two or more processors.
So, only the processors were the multiple things that we had.
But in cluster system,
we have the complete system that is coupled together.
that means two or more individual systems,
that means complete system,
not only the processors but the complete system
are coupled together and they form a cluster.
So, the is what we mean by this cluster system.
So, this system provides high availability.
Now you may have already guessed why it provides high availability.
that is because we have more than one system coupled together.
So even if one fails,
the rest of the system together can take care of the task
that was being accomplished the one that failed.
So you are not having complete system failure even if one or two fail.
Because there are others that can take care of the system.
So, a chance of complete failure is very less.
that is why we say that, it provides high availability.
and even in this clustered system,
this can be structured either asymmetrically or symmetrically.
So, just like our multiprocessor system,
even this clustered system can be structured either
asymmetrically or symmetrically.
so, in asymmetric, what happens?
So in asymmetric, one machine is in hot-stand by mode,
and the others run the applications.
So we know that in asymmetric mode,
what we have is 'master-slave' approach.
So even in this one, one machine
it is in a hot-standby mode, so that is like the master.
then the other systems run the applications.
So, this machine in hot standby mode monitors the other server
So the other systems which are running
And if it finds that some of them fail
if let's say one of them fails
this machine in the hot-stand by mode will take its place.
So, one is monitoring and the rest of them are running.
and then in symmetric mode, two or more hosts run applications
and they monitor each other.
So, unlike this were one is monitoring, here all the systems or the host
are running the application or involved in certain task.
And instead of one system is monitoring everyone,
here all the systems are monitoring each other.
If you think which is better? asymmetric or symmetric?
Obviously it is the symmetric one.
that is because here, all the resources can be shared and used efficiently
as compared to the asymmetric structuring.
So, that was about the clustered system.
So, those were the type of computer systems
based on the number of general purpose processors.
We discussed about the single processor system,
multiprocessor system, and clustered system.
And we also saw the advantages of one over the others.
And we have also seen the types of these systems.
So, I hope you understood these types of computer systems
based on the number of general purpose processors.
I hope this was clear to you.
Thank you for watching and see you in the next one.
5.0 / 5 (0 votes)