| Introduction Syllabus | Introduction : Definition-Relation to Computer System Components - Motivation ~ Message . Passing Systems versus Shared Memory Systems - Primitives for Distributed Communication - Synchronous versus Asynchronous Executions - Design Issues and Challenges; A Model Distributed Computations : A Distributed Program - A Model of Distributed Executions - Models, of Communication Networks - Global State of a Distributed System. _—$$$$$$$—$$ oo Contents . 4.4 Definition see ec eee eeececeven May-22, ----+-- Meets ees Marks 13 1.2 . Relation to Computer System Components ‘ . 1.3 Motivation 1.4 Message-Passing Systems Versus Shared : . a Lrrr—se—SsS Dec.-22, ---++ 20+ bee ee ee Marks 13 1.5 Primitives for Distributed Communication . 7.6. Synchronous versus Asynchronous Executions | 1.7 Design Issues and Challenges-. . 1.8 ~ A Model of Distributed Computation. 1.9 A Model of Distributed Executions 1.10 Models of Communication Networks 7 : 1.11 Global State of Distributed System 12 Two Marks Questions with Answers 14. May-22, --+ A Distributed Program ++ Marks 13t tributed Distributed Computing 1-3 leocaene Distitted Comping 12 Intodton EERE befinition PAU: May-22 | Definition of distributed systems : \ «A distributed system is one in which components located at networked computers communicate and co-ordinate their actions only by-passing messages. A distributed system is col be indi tion of independent entities that co-operate to solve a solved A distributed system is a collection of independent ‘computers that appears to its users a single coherent system. + DS can be characterized as a collection of mostly autonomous processors communicating over a communication network. It having following features and consequences Concurrency : The capacity of the system to handle shared resources can be increased by adding more resources to the network. The p and q are concurrent if either p can happen before q or q can happen before p, thus having interleaving ‘semantics. 2. No global clock : The only communication is by sending messages through a network, Not possible to synchronize many computers on a network and ‘guarantee synchronization over time, thus events are logically ordered. Not possible to have a process that can be aware of a single global Independent failures : The programs may not be able to detect whether the network has failed or has become unusually slow: Running processes may be unaware of other. failures within context. Failed processes may go undetected, Both are due to processes running in isolation. 4, Autonomy and heterogeneity : The processors are loosely coupled in that they have different speeds and each can be running a different OS. o RI Disadvantages of DS 1. Software : Difficult to develop software for distributed systems. lossy transmissions. Easy access also applies to secret data. 4, Absence of global clock upthnst for knowsedge ERE Difference between Parallel Computing and Distributed Computing Parallel computing is typically fine grained with low overhead, 21 Assumed tobe slible Parallel computation » values. shot | excction time, Wee 1. Explain how a parallel systen: ifs rom a distributed sytem. EQ Relation to Computer System Components ‘+ Fig. 121 shows typical distribitted system. iacesor | Mea Fen | aa Processor | Mono Processor | Meany Proceseot | Memory Fig, 4.24 Distributed system In disteibuted com ‘memory and int message passing be TECHNICAL PUBLICATIONS® - an up frKrowedeInrodstion + Databeated lyereel mn ‘The distributed system uses a system design, + Eich computer has memory processing unit and the computers are connected by communication network: All the computers can communicate with each other trough. LAN and WAN. The distributed systein uses a layered architecture to break clown the complexity of system de: ‘+ A distritbuted system is an information-pro of independent computers that cooperate wi networl< in order to achieve a specific objective. * Usually, distributed systems are asynchronous, lock and do not impose any bounds on transfer times. Differences between the various comput they comnmunicate are mostly hidden from users. system that contains a number ne another over a communications they do not use a common peeds or message he wajs in which + Users and applications can interact with a, distributed system in a consistent and uniform. way, regardless of where and Interaction takes place. Each host executes components and operates a distribution middleware. * Middleware enables the components to co-ordinate their activitis..Usérs perceive the system as a single, integrated computing’ fac ‘© A distributed computer system consists of multiple software components that are fon multiple computers, but run as a single system. The computers that are in a distribuated system can be physically close together and connected by networlx or they can be geographically @istant and connected by a wide area network. ‘A distributed system can consist of any itumber of possible configurations, such as mainframes, personal computers, workstations, minicomputers and so on. {EI Motivation 1. Econormics : A collection of microprocessor3 offer a be mainframes. Low price /performance ratio is the cost computing power. 2 Speed : A distributed system may have more tolal computing power than a 3. mainframe. extended through the addition of components, thereby compared fo centralized systems. ruted systems « providing better scala Some applications are inherently distributed eg. a lity + IF one machine crashes, th ey crashes, the syst as a whole can sti and improved reliability. : ‘omputing, power can be added in small increments, Nood of Distributed Systsm ation of the distributed system. The fem “resource” 4s a rather abstact one, but it best characterizes the range of things that ean twefully be shared in a networked computer system, rm + Resources may be the software resources or hardware resources. Printers, diss, (CDROM and data are the example of software and hardware TesOUrCes, Sharing of resource extends from hardivare components such as disks and printers tg software ~ defined entities sich as files, databases and data objects of all Kinds, 7 Se also includes the stream of video frames and audio connection that a motig ‘© phone call represents. A resource manager is a software module that manages :” set of resources of a sributed system are as follows 2. Consistency of replicated data | 4, Reliability ‘5, Concurrent transactions ‘+ Primary requirement of 1. Fault tolerance 3. Security Focus on Resource Sharing | + The fem resource is a rather abstract one, but it best characteris the rap things that can usefully be shared in a networked ‘computer system, am ovd meee . boeing oe on npn ps only be accessed from other computers by comsunication. For effective each resource must be managed by a program called resource manager of communication interface enabling the resource being accessed, manipula" updated consistently and rel ' rely wy + Equipments are shared to reduce cost. Data shared in database or we pp} high-level resources which are more significant to users without regard in server or servers that provide these. | + Types of resources : 1. Hardware resource : Hard disk, printer, camera, scanner 2. Data le, database, web page. 3. Service : Search engine TECHNICAL PUBLICATIONS® - bn ypannst or knowldge TECHNICAL PUBLICATIONS? - an ups fr loweEL noduetion ‘+ Pattems of resource sharing, vary w closely users work together : Users cooperate directly le user's action are determined by the pattem of sharing and the geographic distribution, + For effective sharing, each ufce Must be managed by a progeam that offers a ed resources and presents their functionalities to users and applications. Server is basically storage of resources and it” provides services to the ‘Authenticated clients. It is running program on a networked computer. Server Accepts requests from client and performs a service and responds to. request Example is Apache server and IIS server, ‘The complete interaction between server machine and client machine, from the jent sends is request to when it receives the server's response, is wocation. + called a remote * Resources may be encapsulated’ as objects and accessed by client objects, In this case a client object invokes ¢ method upon a server object. 'tdware and software resource sharing © Examples of hardware resources that can be usefully be shared and examples of their sharing Fdware resources = 1. cpu: a. Computing server : exectites processor-intensive applications for clients. bb. Remote object server : I executes methods on beh o c Worm program : It shares CPU capacity of desktop machine with the loca user 2. Memory : Cache server or faster recently-accessed web pages in its RAM, f access by other local computers server, video on demand server. as X11, allow processes in remote TECHNICAL PUBLICATIONS® . an upvus! fr hnowiedge . EE “=C____ Distabtod Computing Networked printers accept int them witha queuing stem, P* HOM many computers and Software resources 2 Web page : Web servers enable mtple clients to share readonly age content le clients to shate read-write files. Possibilities for software objects a iary and room booking system are examples of less. Shared whiteboard, shared this type. 4. Database : Databases are intended to record the of data, They have been shared ever ‘They include techniques to manage concur definitive state of some related since multi-user computers appeared. rent updates 5 Newsgroup content ; The netnews system makes read-only copies’ of the 'y;posted news items available to clients throughout the Intemet ‘audio stream : Servers can store entire videos on disk and deliver them at Playback speed to multiple clients simultaneously. HE] Message-Passing Systoms Versus Shared Memory Systeme LAU - Dec.22 | Message passing + Two processes commuhicate with each other by passing messages. Message passing is direct and indirect communica Indirect communication uses railbox for sending receiving message from other process. Message passing system requires the synczonzaion and communication between the two process, Message passing used as a method of communication in lhe 4D a yaya Message passing systems come in many forms, Messages sent by a process cin be either fixed or vassbe size. The a provided of message passing is normally a pair of ime, message) 7 (souree.name, message) lve is used for sending a message to dest ‘meetIntroduction Distibuted Computing 1-9 inde Diibted Conres EEE Emulating Message - Passing Systems on a Shared Memory space Systems Seared mene those in which there is common shared address sp ys ‘6 Shared memory systems are srocessors takes place via shared + Shared ass spe 18 divided into two dijint pts snd assigned to ex i i 7 processor. Send and receive operations are implemented by using writing and ata variables and ox estat it tion receiver and sender processor. “ to ss hy competing pans & bls Feading eosin fr sender proce 2 Fo ot ven aachonge information by reading and writing data to the Synchronization primitives lke write and read opettion are controlled by s aed sender/receiver processor. AN)=F; mesage passing canbe emulated bya write by to the maibox and tana rnd by Ff th bon ae jence of communication, as it f Sort cess main pd nn dns at nema pets heh win 3 comput soon 15 message-passing systems are ter than message pang, a8 Mee ss ae lemeted sing syste and ths rope the ceuming sk of Kernel terverton (astm 41. Mlustate the difference between message passing and shared memory process communication mode CEE EE Primitives for Distributed Communication =| Message - passing systems J tat exchange messaging or fae caled message P=sing | EIERI Blocking / Non-blocking, Synchronous ! Asynchronous Primitives ping i stl for saring Mar Sedans feat so that cont * Message send communication primitives is denpted by Sendi wd ot occur ‘communication primitives denoted by Receive(). © | t + Message passing primitive commands \ SEND (msg, dest RECEIVE (sre, bu Send primitives uses two options for sending data : Buflered and unbuffered. In buffered options, user data-is.copied the data gets copied directly from the-user { | smesage psig the eammunication I owertnfen compared fo shared memory techaigue ca passing technique. rece cama ith te changing mesg In unbuffered opts, fer onto the network. ‘The communication of a message betwoen two processes implies some level of Bele, Lele] aces Synchronization between the two processes. Sender and receiver can be beckig E cr nonbocking, Three combinations are possible using blocking and nonblockg 3 Blocking send, blocking receive. Nonblocking send, blocking receive. 1 Blocking send, blocking receive : Both the sender and.recever are thle) . This is called Rendezvous. This combinatia is technique ean be used in eteogensous computers. My r Peale lars i Interconnection network IE I I I eT: Interconnection network mf fe |e dm, i ___ eee ee ‘m0 nv manos? « an uptnst for knowledge oe La e TECHNICAL PUBLICATIONS” an uptnistforknomedyeInteseton ls to be blocked until such message Process that exists to provide a service or et patty is requited to wait, * Processor synchrony + sor synchrony indicates that all the processors ‘execute in lock step with, locks synchronized. KEG synchronous versus Asynchronous Executions Synchronous execution : Main features : A system in which the following bounds are defined 1. Lower and upper bounds on execution time of processes can be set. 2. ‘Transmitted messages are received within a known bounded time. 3. Drift rates between local clocks have a known bound. Important consequences : 1. Ina synchronous distributed system there is a notion-of global physical time (wth a known relative precision depending on the drift rate) 2. Only synchronous distributed systems have a predictable behaviour in terms of timing. Only such systems can be used for hard real-time applications. 3. In a synchronous distributed system it is possible and safe to use timeouts in order to detect failures of a process or communication link. + Ikis difficult and costly to implement synchronous distributed systems. ‘Asynchronous executions : ‘© Many distributed systems (including those on the Internet) are asynchronous. ‘+ No bound-on process execution time i.e. nothing can be assumed about speed, load, and rel ry of computers. + No bound-on message transmission delays ie. nothing can be assumed about speed, load, reliabil ‘+ No bounds on of interconnections. tes between local clocks. Important consequences : 1. In an asynchronous distributed system there is no global physical time, Rea: can be only in terms of logical time. be only in terms of logical ime TECHNICAL PUBLICATIONS® - on uptirust for krone Distibuod Computing 2. Asynchronous distributed systems are uny 3. No timeouts can be used, ba + Asynchronous systems are widely and success Synchronous execution Syrcrenus execution means te fat ina program must finish rocesing before executing the next and upper bounds on execution processes can be st smite messages ae recived hin known bounded time betwen local clocks have a a EEA Design Issues and Challenges EEEEH chatienges trom system Perspective * Communication mechanisms : ‘This task involves designing appropriate mechanism for communication among the processes in the network. For example Remote Procedure Call (RPC), Remote Object Invocation (RON, message-oriented 1. steaoriented communication, = : hud «omer, atin (eet fiesmetean dA ‘eal dak, a rocesses : Issue involved are tode migra, progest/hread! management at tints and servers, design of software and mt asent + Naming Easy to use identifiers needed to locate resources and processes transparently and sea Synchronization ynchronization or coordination among, the exclusion is the cassical example of manly other forms of synchronization, such as leader election are also needed. Data storage and access : Various schemes for data storage, searching and lookup should be fast and scalable across network. Revisit fil system design TECHRUCAL PUBLICATIONS” - an upto kontoPitromdcomery (key gen the various method used to provide socuri EERE chatienges «Dagan he dsb be overcome in onder to Cieburd ten ae 08 1. Heterogeneity 2 Openness soupy 8 Firatndg 7 Tansey : ERE tewozeiy 3) : shen ta cm, ey eis st avi poner spec ik oP iding ei flowing of faisest applies to al the fllowing one aae LAN, aless network, stellt link Computer hardware devices i Laptop, computer, mobiles phones, does not come for free. Some challenges need to ideal systems. Design issues and challenges of 3, Security 6. Concurrency Operating systems : Linus, UNEX, Windows Programming langusge® ¢ C/CH Java, PHP - es of stare developers, designers, system managers vee eral representations of data in the system. This mi for integers, byte streams, floating point numbers, 1 2 3. 4 5 «There may be many di include different representations and character sets. e Most of the data can be marshaled from one system to another with significance. Attempts © Provide a universal canonical form of informat ferogeneous components implies the construction of tems, Ind distributed system, heterogeneity is almost unavoidable, nents MAY require different implementation technologies. iddleware . idleware is a softvare layer that provides a progr heteogeneily of the underlying, platform. Eg., CORBA, DCOM, ming abstraction as’ well 13 ht, rT aspects by providing uniform, stag, | the application developers and integrators, so tay ions can be easly composed, reused, ported, and made to inferopea, Midaleware services provide common services to perfor various general pup | functions, Fig.1.7.1 shows position of. middleware. Middleware software resides above the network THT | | and below the application AP or sandarize,hgh-evel senicas software + Mobile code is used to refer to So Miktoware ces ft | Ty transferred from one computer to ‘another and run at the destination _[ Dicibuléd heterogeneous hardware nods | computer. Example of mobile code is Java applets. Code suitable for running on one computer is not necessarily suitable for running | fon another because executable programs are normally specific both to the | {instruction sét and to the host operating system. Middleware should make the | ‘network transparent to the applications and end users. Users and applications should be able to perform the same operations across the | network that they can perform locally. Middleware should hide the details of ‘computing hardware, OS, software components across networks, EEF openness * Openness means that the system can be easily extended and modified. Openness refers to the ability to plug and play. You can, in theory, have two e serics tat follow the same interface contrat, and interchange one with the | other. | 5 Applicaton sotware Fig. 4.74 Position of middleware *, The integration of new components means that they have to be able to | Guumunicate with some-of the components that already exist in the system Openness and distribution are related. Distributed system components achieve | ‘openness by communicating using well-defined interfaces, XT the welldefined interfaces fora system are published, itis easier for developers to add new features or replace sub-systems in the future, Open systems can easily be extended and modified. New components can integrated with existing components TECANCAL PUBLICATIONS® «an up-ust fr krowiedge TECHNICAL PUBLICATIONS? . an up-tist fr knowledgeIntec need to be published. New components have to be with existing components sm needs to have a stable architecture so that new components can be easily integrated while preserving previous investments. + An open distributed system offers services according to standard rules that describe the syntax and semantics of those services. EET security + Security becomes even more important in a distributed system. Authentication, authorization, digital signatures, non-repudiation, encryption, and privacy become major isstes in the distributed system. +The four basic goals of a security system are to protect information, to detect an intrusion, to confine the security breach, and to repair the damage and retum the system to a known stable and secure state. «Secu for information resources has thee componen’s: Proton against dilosire to unautorized individual, eg ACL in uni file system. Integrity : Protection against alteration or corruption, eg. checksum. jon against interference witht the means to access the seg. Denial of service. + Encryption provides protection of shared resources, keeps sensitive information Security challenges that are not yet fully met : 2. Security of mobile code. + A denialof-cervice attack is an attempt fo make a computer or network resource le to its intended users. mobile code : Mobile code systems are conceived to operate in large here networks are composed of heterogeneous hosls, managed by ‘with different levels of trust and connected by links with TECHNICAL PUBLICATIONS® - an up-tust or knowldgn Disrbuted Computing 15 Inotucion * Mobile code systems address a wide range of needs and requirements, such as service customization, dynamic extension of application functionality, autonomy, fault tolerance and support for disconnected operations. REA scataitty is said to be scalable if it can handle the addition of users and resources without suffering noticeable loss of performance or increase in administrative complexity. “The ability to accommodate any growth in the future be it expected oF not Distributed system architectures achieve scalability through employing more than ‘one host. Distributed systems can be scalable because additonal computers can be added in order to host additional components aling with large numbers of machines, users, tasks, Dealing with geometric distribution and mobility 4. Avoiding performance bottlenecks. Controlling the cost of physical resources Le. servers and users. linear structures and save time for access structured data. Preventing software resources running out : Internet 32 bits addresses run out soon. 128 bits one gives extra space in messages. = ‘Avoiding performance bottlenecks : DNS name table was kept in a single master partitioning between servers. large directories and large numbers of files while s providing 1/0 performance. Google file system aims at eliciently and relably ‘managing many extremely large using commodity hardware. iques such as replica jonous processing hel s for many clients, caching and cache memory management TECHNICAL PUBLICATIONS® - an ups fr knowledgefrom heterogeneous hardware and software. Openness is sions and improvements of distributed systems. Detaled sponents need to be published. New components have to be components can be ‘+ An open distributed system offers services according to standard rules that describe the syntax and semantics of those services Security ‘+ Security becomes even more important in a distributed system, Authentication, authorization, digital signatures, non-repudiation, encryption, and privacy become major isies inthe distributed system. + The four basic goals of a security system are to protect information, to detect an intrusion, to confine the security breach, and to repair the damage and return the system to a known stable and secure state Security for information resources has three components : f shared resources, keeps sensitive information ges that are not yet fully met secret when transmitted. Security 1. Denial of service attacks mobile code. + A denial-of-service attack is an attempt to make a computer or network resource unavailable to its intended users. «Security of mobile code : Mobile code systems are conceived to operate in large scale settings where networks are composed of heterogeneous hosts, managed by ities with different levels of trust and connected by links with different bandwidths. Mobile code systems address a wide range of needs and requirements, a service customization, dynamic extension of application functional * ful toeranee and supp fr dsconneced opens iy | A oem ade a iit can he ean of wad en sit ig» no I of pms ns eee | coat The aly © ace ay goth ef be ee Fw Dit yen hese see salty gh eects a one hot Datel se an be ai bee sehen ee ae Siete aa 2th ation : Dealing with geometric di 3. In administration : Addressing data passing through different regions gl ownership. ‘+ The design of scalable distributed systems presents the following challenges, + Controlling the c ent | cence from running out. vance bottlenecks. of physical resources i. servers and users. Controlling the performance loss : DNS hierarchic structures scale better fur linear structures and save time for access structured data, Preventing software resources running out Internet 32 bits addresses nin ot 500n. 126 bits one gives extra space in mi : "formance boitlenecks : DNS name table was kept in a singe ait file partitioning between servers. rectories and large numbers of Providing 1/0 performance. Google file system managing many extremely large mn, caching, and cache memory marge and asynchronous processing help to achieve scalability. a TECHNICAL PUBLICATIONS® - an upstrus for knowledge TECHNICAL PUBLICATIONS® - an upsust fr knowledgeunwerming scaly ehniques kdag comenication Bramples would be asynchronous as csion. as well as pushing code down to clents (eg. Java applets and Jjuascript) + purbution + Taking a component, spliting into smaller pats, and subsequently penfing ther across the system, fencies 4k Rplation + Replicating components increases availabilty, helps balance the load lading to better performance, helps hide latencies for geograpl nated sens. Caching isa spedial form of replication. [EBB rae Hanating "en faults occur in hardware or software, programs may produce incorrect mis or they, may slop before they have completed the intended computatin our hanalling is dificult in distbuted systems because failure is paral Le fone components fail wile ther continue to function + thnvare, softwvare and networks are not fe of failures. Operations th aan ihe presence of fuls are refered to a5 faulttoleant. p19 ted systems can maintain avaibity even at lor levels of hardware «+ Distut Jastvare /network reliability. 1s Falt tolerance, that should not involv ty covery and replication of components fe users or system administrators, is achieved «+ Techniques for dealing with failures 1. Detecting, failures 2. Masking failures 4 Tolerating failures 4. Recovering, from failures 5. Redundancy: ing failures = For exampk Detection of failure in remote cras Not all failures are detected but some of the failures can be le: Corrupted data from file is detected by using checksum. hed server in the internet is hard to detect «Masking, failures + Failures are hidden or made Tess sever. Examples of hiding 2. Files can be ted on different place. Distibuted Computing ‘ Intedton Tolerating the intemet, client 7 : which generally invol # an be designed to tolerate failures ‘he users tolerating them as wel Recovery om flues: erry ne ty = Recon nls he dan of pin dat cat bdo nk ae a ste Redundancy : Services can be made to tolera mete to tolerate failures. Let us consider following ‘There should be at least 2 routes between any 2 routers in the internet In DNS, every name table is replicated in a st 2 servers. A dase canbe replied in sever serves eed sever seve. Clns ae sliced fom failure one to working one. i “ ee HEED concurrency Components in distributed systems are executed in concurrent processes. Fig. L72 shows concept Tere of concurrency. Teak ze environment ‘Any object that represents a shared resource in 4 distributed system must be responsible for ensuring that operates comely in a concurrent environment. Fig. 1.72 Concurrency FREE Transparency Distributed systems should be perceived by users and application programmers a3 1 whole rather than as a collection of cooperating components. Transparency is @ key issue within distributed system, Hide different aspects of distribution from the client. Its the ultimate goal of many distributed systems. It can be achieved by providing lower-level services. The client uses these services instead of hard coding the information. The service layer provides a service with cettain Quality of Service. _ TECHNICAL PUBLICATIONS® - an up-tus! fer koowedge TECHNICAL PUBLICATIONS - an up-to tnofedoowai eescan + Tonsparency is Important goal, but has to be considered together with all rirements and with respect to particular demands. resources need not know the location of reve or update data from any si em {othe site or sites related to the Tocation transparency : User usin the resource. Any request automatically forwarded by the sys processing request For example: URL Access. transparenéy + Using identical operations to acc resources, eg, Hyperlink in web page. Concurrency iransparency + Several processed operate concurrently wing shared resources without interference with between them. tances of resources to be used 10 ‘knowledge of the replicas by users ss local and remote Replication transparency. + Multiple inst ity and performance without eg. we cache. rations to complete their tasks despite i ponents, & or application programmers, Failure transparency : Users and applic the failure of hardware and software comy Mobility transparency : Movement of resources and d without affecting the operation of users and programs, €:8 transparency : Allows the system to be reconfigured (0 improve ts within a system ‘mobile phone. Performance performance as loads vary. the system and applications to expand in sale scaling transparency : Allows jibe atte te tructure or the application algorithms. wwithoutt change to the system st + Advantages of transparency * + Easier for the user: a. Doesn't have to bother with system topography b. «+ Easier for the programmer : Doesn't have fo know about changes: Easier to understand. Doesnn't have to bother with system topography” b. Doesn't have to know about changes. ¢ Easier to understand. pisadvantages: of transparency : 9. Optimization cannot be done by programmer oF Use b, Strange behavior Underlying system ca ing system be very complete. Sil es ee ICAL PUBLICATIONS® - an up-to fer Konfodgo rec istrinted Computing 119 ERED 40 Ltn tion of Distributed Computing and Challenges te Mobile system ‘The portability of the devices, such as laptop computers, PDA, mobile phon refrigerators, together with their ability to connect conveniently to networks different places, makes mobile computing possible. ‘computing. is the hamessing of many small cheap computation devices that are present in use’s physical environments, including the home, offi and elsewhere, Mobile devices are 1, Laptop computers. 2. Handheld devices, including PDAs, cell phones, pagers; video cameras « 3, Wearable devices, such as smart watches. 4. Devices embedded in appliances such as washing, machines, hii systems, c ‘Mobile computing (nomadic computing) 1. People can still access resources while.he is on the move or visiting pl ‘other than their usual environment. 2. Locatior-aware computing : Utilize resources that are conveniently nearby. Ubiquitous computing (pervas 1. The hamessing of many smal in user's physical environments, suding the home, office and elsewhere 2. It benefits users while they remain in a single environment such as home. Fig. 1.73 shows the portable and handheld devices in a distributed system (Gce Fig. 173 on next page) Mobile and ubi architecture for mobi i$ computing raise significant system issues presents computing. User has access to three forms of wireless connection 1, A aplop is connected to hosts wireless LAN. 2. A mobile phone is connected to internet using WAP via a gateway. 3, A digital camera is connected to a printer over an infra-red link Se eee TECHNICAL PUBLIGATIONS® - an up-thust for rowedeited Computing betrocton Fig. 1.7.3 Portable devices in DS 2. Pervasive computing fas become a pervasive resource and devices ‘can be connected at any Mime and ay pace The mode Teme i collection of various compiler Rakion es, Example of network includes a wide MAX, Bluetooth + Computer network + Big. 1.74 shows th + Progeams running on the sssion capacity, employing vue abo proves tine lene user can watch User can dowel video conference. TV, play online games and do the vi sean nae 2 pale sr Hat consiat of many intfinked Jo area. ne WAN. «Intranet is composed of connections Rou rise, It may contained sithin an enterp spoorks and also wee Teased tines in the tocal area networks linked by backbone boundary that can be intranet. | by preventing unautborized messages kaving 4) ered messages. The cst of software installation ang reduce by the use of system architectures such as network ‘0 comptes and thin clients | 43. Multimedia system . | + Digital multimedia : Computer-contoled integration of tex, graphics stil images moving imation, sound, and any other medium. All thece data typq ted, essed digitally. — + Multimedia streams are said to be time-based because tinied data elements { audio and define: the content of the stream. The systems ty applications need to preserve the timing when they hand é ‘media types tothe user and optionally also to share the me of use presentation bane wo dees ype ey inkrton fr cont! noma wl Web casting 4 * Web casting is an application of distributed multimedia technology. It continuous media over the intemet. A web ming media teal ‘Yo. many simian — method of obtaining informal content to your computa 84 { TECHNICAL PUBLICATIONS® « an up-thast for kowedye a aa erInduction rived Conoutind (ii desig ues and legit stl ste fo ma system perspec DE HED A Model of Distributed Computations : A Distributed Program sed of a set of "n” asynchronous Processes like py the communication 1. Discuss th « Distbuted program is compos par Par + Pir Pa that communicate work. ae = cach process is runing on a iferent prosesoh ‘The passing messages method >, to process py and let my denote = te and unpredictable by message pasing over + Here we assume that, processes communicate only By wn denote the channel fro" Process Py ‘The communication delay i fin a bal clock. Process execution ad MESSAGe soba execute an ston sportanestsy ad 9 y For the delivery of the message (0 + Let Cy message sent by Pi © Pj Also, procesies do not share a glbal ier are asynchronous. ~ 8 process process sending a message does 7 complete. EEDA Model of Distributed Executions ‘Process execution means, sequerta) regone of process ae of Aree TYPES receive events. message 72 yt proces wich sing Be ee i = ee cent Aft ee ananassae he See ss messages from ater Processes: en ea cet te BS — a gem, sen) denotes send events and reel) 1 eng sk) dO Tac me gs tt eer on ish he me ve fe poe receive event changes the is received. + sig ron whch the snes = s state of agram for distributed execu «Fig. 191 shows spacesiming 4 : Nee ey action. Actions are atomic. send events and jon of its val events, Message Fig. 19.4 —_—— an apt fr tC =e OBLIGATIONS enote its receive That sends the message and ceives the message and the Disibuted Computing ae Introduction + Th end an the tv et yt oof ns and establish causal dependency from the sender proces to the + Inthe figure, for proces event san itera even 25 ea red Causal precelonceeaton) betiveen processes receiver process w second event is a message send event, the thind the fourth event is a message receive event. * Ordering of eveitls“for a single process is 9 ile process is simple: they are ordered by their Send and reve evens sig the ow of in low of infomation between process exablsh aul precedence beveen evens the sener and meanest For any two events e; and e,¢; > 6; denotes the fact that event e; does not di transitively dependent on event e, eos ly dependent on event e,. That is, event e; does not causally vs. Physical concurrency + Ina distributed computation, two ev ton, two events are logically concurrent if and only i they do not causally affect each other. see * sical concureny, om the oer hand has coolio thatthe events Gur a e same instant 7 ssor speed_and message vents could have very W coincided i physi Whether a set of logically concurrent events coincide in the physical ine o no does not change the outcome of the computation. ' + Therefore, eventhough ast of logically concurrent evens may not have cured atthe une tat pos ysl ie an sume ate een ae the same instant in physical time. = = send (taj) then ree (11 yy) > ree (my). + This property ensures that caus destination are delivered in an order tha Causally ordered delivery of messages implies es the design o i same ally related messages destined tO oy ae consistent with their causality relation- IFO message delivery. f distributed algorithms | * Causal ordering model considerably simpli because it provides a built-in synchronization. Il Global State of Distributed System jon is of all Definition ; "The global state of a distributed computation is the set of local pa : f a individual processes involved in the computation plus the state of the communication ch . Requiroments of global states « Fig. 1.11.1 shows detecting global properties. 2 message a. Garb; object ‘age collection Hee — CC —- ‘garbage object ©. Termination Fig. 1.14.4 . Distributed garbage collection : An object is considered to be garbage if there are no longer any references to it anywhere in the distributed system. It is based on b reference counting and should include the state of communication channels. 2. Distributed deadlock detection : It occurs when each of collection of processes waits for another process to send it a message and look for “waits-for” relationship. Distributed termination detection : Look for state. in which all processes are Passive. 4. Distributed debugging : Need collect values of distributed variables at the same time. TECHNICAL PUBLICATIONS® - an up-thrust for knowledge BI My