Agenda In this chapter, we take a closer look at: The role of - PDF document

Lecturer: Hadi Salimi Lecturer: Hadi Salimi Distributed Systems Lab, School of Computer Engineering I Iran University of Science and Technology, U i i f S i d T h l hsalimi@iust.ac.ir Agenda � In this chapter, we take a closer look at: � The role of processes in distributed systems � A brief introduction to virtualization technologies. � Other process ‐ related concepts like migraion. O h l d lik i i

Threads � The granularity of processes in a distributed system is not sufficient. � There is need for a more finer element inside processes. � Using threads, can bring different advantages and di disadvantages. d Traditional Thread Usages � The spreadsheet example. � From an IPC point of view. p � Threads are more proper in some cases. � Are web browsers a good example?

Thread Implementation � Threads could be implemented as: � User mode threads � Easier for implementation � Low overhead � A thread blocks leads to process block � A thread blocks leads to process block � Kernel mode threads � Hard to implement � A blocked thread does not block the whole process � Hybrid threads (LWPs) � Uses a hybrid approach as depicted. U h b id h d i d Hybrid Thread Implementation � Threads are supported on both kernel and user modes.

Thread in Distributed Systems � They could avoid a process from being blocked when a blocked system call has happened. � Threads could be used at: � Client Processes � Server Processes Multithreaded Clients � A multi ‐ threaded browser: � Performs communication and page rendering simultaneously. � Could have multiple connections at the same time. � This could be more useful when the browser gets the page � This could be more useful when the browser gets the page form a replicated server.

Multithreaded Servers � As the first example consider a file server: � In the case of a blocked thread, other thread could get , g client requests. Virtualization � The concept of virtualization: � To present a concept in a new form � This technique has been applied for many decades. � Role of virtualization if distributed systems: � Legacy support � Easier migration E i i ti � More flexibility � More flexibility

Virtualization � This figure shows the role of interacting interfaces on a computer system. Architecture of VMs � Interfaces at different levels � An interface between the hardware and software consisting g of machine instructions � that can be invoked by any program. � An interface between the hardware and software, consisting of machine instructions � that can be invoked only by privileged programs, such as an operating system.

Architecture of VMs � Interfaces at different levels � An interface consisting of system calls as offered by an g y y operating system. � An interface consisting of library calls � generally forming what is known as an application programming interface (API) programming interface (API). � In many cases, the aforementioned system calls are hidden by an API. y Architecture of VMs � Interface stack on a computer system.

Architecture of VMs � A VM that can provide a virtual environment for an application. Architecture of VMs � Virtualizing the hardware for an operating sytem using a Virtual Machine Monitor

Client ‐ Server Computing � Client � Networked User Interfaces � The X Window System � Compound Document � Client side software for distributed transparency � Client ‐ side software for distributed transparency � Server � General Design Issues � General Design Issues � Server Clusters � Distributed Servers � Managing Server Clusters � PlanetLab Networked User Interfaces � Two kind of clients: � Application ‐ specific protocols for clients pp p p � Example: X Window � General ‐ purpose protocols like TCP/IP

X Window Systems � The basic organization of the X Window System Compound Documents � Today modern documents are composed on many different parts, each of which belongs to an special application. � As an example, a word processor document could contain: A l d d ld � Text, graphic, movie, charts, database tables, etc. � Like the X Window system most of the computations are � Like the X Window system, most of the computations are done at server side.

Client ‐ Server Computing � Client � Networked User Interfaces � The X Window System � Compound Document � Client side software for distributed transparency � Client ‐ side software for distributed transparency � Server � General Design Issues � General Design Issues � Server Clusters � Distributed Servers � Managing Server Clusters � PlanetLab Clients and Distribution Transparency � The client software could contain small processing elements as well. � The small elements could use to bring different types of d distribution transparency such as: b h � Access transparency A t � Replication transparency � Replication transparency � Failure transparency p y

Clients and Replication Transparency � A client can easily provide replication transparency through intercepting requests. Client ‐ Server Computing � Client � Networked User Interfaces � The X Window System � Compound Document � Client side software for distributed transparency � Client ‐ side software for distributed transparency � Server � General Design Issues � General Design Issues � Server Clusters � Distributed Servers � Managing Server Clusters � PlanetLab

Server Design Issues � A Server is a process the implements a specific service on behalf of a group of clients. � Servers could process incoming requests or easily send it to another process. (e.g. using fork) h ( f k) � Clients and Servers need an end point to communicate, so Cli d S d d i i how to make it? � Publishing the endpoint globally � Publishing the endpoint globally � An agreement between client and server � Using a Naming Service g g End Point Table � A case in which a client could find the end point by means of a special daemon server, equipped with an end point table.

Many Servers ‐ On End Point � Having a server for each end point may waste the resources. � This can be prevented by having a Super ‐ Server acts as a multiplexer multiplexer. State Manipulation � A stateless server does not keep information on the state of its clients and can change its own state without having to inform any client. � A simple time server � A file download server A fil d l d � When the request processes, the server completely forgets about the client. � Some unimportant information such as log files may be kept in stateless servers. � But loosing this information does not lead to any critical disasters. � These kind of states are called soft ‐ states � These kind of states are called soft ‐ states.

Statefull Servers � Statefull servers keep persistent information about clients. � As an example consider a document management server. � This server needs to keep track of the files or documents locked by each client for update. l k d b h li t f d t � Benefits: � Are usually faster � Are usually faster � Disadvantages: � A server crash could not be recovered easily A server crash could not be recovered easily. State Management � Statefull servers are usually persist their data on data layer. � Other info kept on processing layer, such as session info are usually soft state.

Cluster Computing � Three different layers of server cluster. Code Migration � There are cases in which passing programs in execution instead of their data simplifies the design of distributed systems. � Moving a process to a remote machine is a time h consuming activity. � This is usually done in support of load balancing. � The key idea behind code migration decisions is the ratio of communication on computation. p

Code Migration (Cont.) � As an example, if a client application needs to process large amounts of data from a database, it would be more efficient if the application could be moved near the database. � In a similar case parts of the server could be used moved � In a similar case, parts of the server could be used moved into client side. � Any examples? y p Code Migration (Cont.) � Code migration is also beneficial for decomposing an application into parts and distribute them on the system. � The principle of The principle of dynamically configuring a client to communicate to a server. The client first fetches the first fetches the necessary software, and then invokes the server.

Code Migration (Cont.) � From a specific viewpoint, a process could be viewed as a triple: � Code segment: the part that contains executing code. � Resource segment the part that keeps references to � Resource segment: the part that keeps references to resources. � Execution segment: the part that keeps execution state. Code Migration (Cont.) � Weak mobility: � Only the code segments migrates. � The migrated program should be started from beginning. � Example: Applets � Strong mobility: � Code and execution segments could be moved. � The process could be paused, migrated and then restarted. Th ld b d i t d d th t t d