Distributed processing and parallel processing are two different architectures used to improve parallel processing speed.
Parallel processing utilizes multiple functional units or multipleprocessorWorking simultaneously to improve system performance or reliabilitycomputer systemThis kind of system at least includes instruction level or above parallelism.
Distributed processing is to connect multiple computers in different locations, with different functions, or with different data through communication networksControl systemUnder the unified management and control ofinformation processingTask'scomputer system。
The research and development of parallel processing systems involve computing theory, algorithms,ArchitectureThere are many aspects of software and hardware, but it has a close relationship with distributed processing systems. With the development of communication technology, the boundaries between the two become increasingly blurred.In a broad sense, distributed processing can also be considered as aparallel processing Form.The distributed processing system connects multiple computers at different locations or with different functions or with different data through communication networksControl systemUnder the unified management and control ofinformation processingTask'scomputer system。It is generally believed that the close coupling concentrated in the same cabinet or the same placeMultiprocessor systemOr massively parallel processing system is a parallel processing system, while LAN orWANConnectedcomputer systemIt is a distributed processing system.The parallel operating system in loosely coupled parallel computers is sometimes called distributed processing system.
The distributed processing system includes six elements: hardware, control system, interface system, data, application program and people.The control system includesDistributed operating system，Distributed databaseAnd communication protocol.
In short, distributed processing means that multiple connected computers undertake different parts of the same task, run simultaneously under the control of people, and complete the same task together.
By multiple autonomous, interconnectedinformation processing system , under the coordination of an advanced operating system.
Using network technology, many minicomputers or microcomputers can be connected into a high-performance computer system, making it capable of solving complex problems.[1]
developing process
Announce
edit
Schematic diagram of distributed processing
Before the advent of computer networks in the 1970s, information processing was centralized or decentralized.Centralized processing transmits all information, including remote information, to a unifiedinformation processing system For processing.Centralized processing can achieve the goal of high utilization rate of equipment and ensure thatprocess informationIntegrity and validity of.A company with many factories adopts centralizedinformation processingThe best results of the whole company can be obtained. Instead of decentralized processing, the optimization of each factory is considered.The equipment required for centralized processing is much more complex than that used for decentralized processing, but it can be shared by most factories to avoid duplication of equipment.The biggest advantage of decentralized processing is that it is simple and can extract data locally, transform the format and process, and finally output the decision.It is very difficult and expensive to concentrate these efforts in a unified program.The greater the degree of decentralization, the greater the possibility of meeting the department's special information processing requirements, but the greater the difficulty of unifying and controlling the information flow.Due to the increasing proportion of data requiring local processing at the data generation site, the shortening of response time and the impossibility of unlimited enhancement of computer processing capacity, computers set up in a decentralized manner are often combined into a computer network to form a distributed processing system.This can not only overcome the shortcomings of decentralized processing, but also avoid the difficulties of centralization (see figure).Since the 1970s, as computers have been increasingly widely used in various fields, national and international computer networks have emergedinformation processing system Its function and structure are becoming more and more complex.Distributed processing disperses equipment from centralized processing system, so that information can be obtained directly from information sources and processed in coordination with each other.This processing method meets the needs of the development of computer networks.[2]
framework
Announce
edit
commonlyinformation processing system The hardware, control point and database of the.From the perspective of hardware composition, it belongs todistributed system There are two types of: ① The system consists of multipleprocessorComposition, but there is a unified input and output system; ②manycomputer system, with multiple input/output systems.There are three types of distributed processing according to the setting mode of control points: ① multiple systems cooperate to execute a task at the sub task level; ②Multiple identicalControl systemCooperate to execute the same task; ③Multiple different control systems perform the same task together.There are three types of distributed databases: ① only some databases have replicas in the primary node; ②Only some databases have directories in the main node; ③All databases are stored in a decentralized manner without a primary node.
Distributed processing systems must have the ability to dynamically combine into systems oriented to different service objects in a short time.The system is transparent to users. Users only need to specify what the system does without pointing out which component can provide this service.The components of the system are autonomous, but not anarchic. They are coordinated by the advanced operating system according to a master plan.Not all hosts in a computer network are distributed.If such systems do not have the ability to dynamically combine and reassign tasks, they are still centralized.Advanced operating system is the key to distributed processing.staydistributed system Complete information is no longer used in. Each component provides its own status information. Advanced operating systems coordinate tasks and reallocate resources based on these status information[1]There are hierarchical relationships but autonomy.
Typical applications
Announce
edit
Distributed splicing processor
1、Distributed splicing processor——It is a newly launched distributed node computer image processing system featuring networking, digitalization, high reliability, high real-time, high flexibility, high scalability and original graphic splicing synchronization technology (ESYNC). It is also a large screen display wall easy splicing system developed for the complex and changeable visual information display application needs of many industries.
2、Distributed splicing processor——Using the network distributed architecture design, it is easy to realize the multi-level network segment subsystem display wall data service information to securely and reliably interconnect the interactive command and dispatching decisions based on.
That is, in the network environment, the signal input node machine can optimize the grouping of distributed signal sources according to different locations, different signal types, different properties, etc;The display output node machine can be combined and spliced according to different locations, different buildings and different occasions, and can be customized separately. The original signal source images can be displayed on a single screen, across screens and across walls, and expanded to cross floors, across locations and across display walls,Realize signal source sharing display and remote call display between large screen systems of multiple command and dispatching center/control room platforms.
Solve the synchronous display of full screen information between the display wall of the video dispatching command center and the display wall of the decision-making room, the analysis and research room, the conference room, etc., which is conducive to the decision-making command, dispatching, and the simulation of performances on the same wall of the leadership of the decision-making room, and helps the control center personnel to accept instructions and make accurate judgments. At the same time, the front-end signal system, display system, control systemDecision making agencies dispatch and backup each other.
3、Distributed splicing processor——Due to the network distributed architecture, based on the CATe5 and above network cable transmission, the easy splicing system wiring is more convenient and simple, and the transmission distance is longer;The large screen display wall does not need expensive high quality DVI/HDMI/VGA long cable, optical fiber or the application scheme of adding extender mode, and the construction cost of the easy splicing system is greatly reduced. At the same time, it is different from other types of large screen splicer systems.
Distributed processing system
Announce
edit
Distributed processing system and parallel processing system are two types of computer architecture.Parallel processing system is a computer system that uses multiple functional units or processors to work at the same time to improve system performance or reliability. This system includes at least instruction level or above parallelism.The research and development of parallel processing system involves many aspects such as computing theory, algorithm, architecture, software and hardware, but it is closely related to distributed processing systems. With the development of communication technology, the boundaries between the two are becoming increasingly blurred.In a broad sense, distributed processing can also be considered as a form of parallel processing.The distributed processing system is a computer system that connects multiple computers at different locations or with different functions or with different data with communication networks, and under the unified management and control of the control system, completes information processing tasks in a coordinated manner.It is generally believed that tightly coupled multiprocessor systems or massively parallel processing systems concentrated in the same cabinet or at the same location are parallel processing systems, while computer systems connected by LAN or WAN are distributed processing systems.The parallel operating system in loosely coupled parallel computers is sometimes called distributed processing system.[3]
Related differences
Announce
edit
The distributed processing system includes six elements: hardware, control system, interface system, data, application program and people.The control system includes distributed operating system, distributed database and communication protocol.Distributed computing environment is a software environment for computing and information processing on a multi computer system with multiple address spaces.Distributed software system is a software system that supports distributed processing, including distributed operating system, distributed programming language and its compiling system, distributed file system and distributed database system.CORBA and COM+are some technologies for designing distributed software systems.Generally speaking (popular is not very scientific, you can refer to the above statement), distributed processing is that multiple connected computers undertake different parts of the same task, run simultaneously under the control of people, and jointly complete the same task
