distributed

• Distributed system


  A collection of independent computers that appears to its users as a single coherent system
• Distributed system


  • Consists of autonomous components (computers)
  • Users think they are dealing with a single system
  • Components need to collaborate
• No assumptions are made concerning the type of computers in a distributed system

  
• Characteristics of distributed systems

  • Differences between computers and communication are hidden from users
  • Internal organization is hidden from users
  • Users and applications can interact in a consistent and uniform way
  • Relatively easy to expand or scale
  • Continuously available, with parts being replaced or fixed without users noticing
• Middleware


  A layer of software that is logically placed between users/applications and operating systems/communication facilities

  
• Distributed systems are sometimes called middleware

  
• Making resources accessible


  The main goal of a distributed system is to make it easy for users and applications to access remote resources and share them in a controlled and efficient way
• Resources can include printers, computers, storage, data, files, web pages, networks, etc.
• Connectivity of the Internet is leading to virtual organizations where geographically dispersed groups collaborate using groupware
• Increased connectivity can also lead to security and privacy issues, such as tracking user preferences and unwanted communication (spam)
• Distribution transparency


  Hiding the fact that a distributed system's processes and resources are physically distributed across multiple computers

  
• Types of transparency in distributed systems

  • Access transparency
  • Location transparency
  • Migration transparency
  • Relocation transparency
  • Replication transparency
  • Concurrency transparency

  
• Attempting to completely hide all distribution aspects from users is not always a good idea
• There is a trade-off between transparency and performance, consistency, and comprehensibility
• Openness


  A distributed system offers services according to standard rules that describe the syntax and semantics of those services
• Interface definitions in an IDL capture the syntax but not the semantics of services
• Extensibility


  An open distributed system should be easy to configure out of different components, and easy to add new components or replace existing ones
• Scalability


  A distributed system can scale in size (more users/resources), geography (users/resources far apart), and administration (spanning multiple organizations)
• Unfortunately, a system that is scalable in one dimension may not be scalable in another
• Attaining flexibility in distributed systems is easier said than done
• Scalability

  • Scalable with respect to size
  • Geographically scalable
  • Administratively scalable
• A system that is scalable in one or more dimensions often exhibits some loss of performance as the system scales up
• Centralized services


  Services implemented by means of only a single server running on a specific machine in the distributed system
• Centralized services can become a bottleneck as the number of users and applications grows
• Centralized data


  Data stored in a single database
• Centralized data can saturate communication lines
• Centralized algorithms


  Algorithms that operate by collecting information from all the sites, sends it to a single machine for processing, and then distributes the results
• Decentralized algorithms


  • Operate without collecting information from all sites
  • Do not send information to a single machine for processing
  • Do not distribute results from a single machine
  • Take into account lack of exact clock synchronization
• Communication in wide-area networks is inherently unreliable and virtually always point-to-point, unlike local-area networks which provide highly reliable communication facilities based on broadcasting
• Geographical scalability is limited by centralized components due to performance and reliability problems resulting from wide-area communication
• Scaling a distributed system across multiple, independent administrative domains is a difficult and often open question due to conflicting policies with respect to resource usage, management, and security
• Scaling techniques


  • Hiding communication latencies
  • Distribution
  • Replication
• Hiding communication latencies


  Avoiding waiting for responses to remote service requests as much as possible, e.g. using asynchronous communication or moving computation to the client
• Distribution


  Taking a component, splitting it into smaller parts, and spreading those parts across the system
• Examples of distribution


  • Internet Domain Name System (DNS)
  • World Wide Web

  
• Replication


  Making copies of a component across a distributed system to increase availability and balance load
• Caching


  Making a copy of a resource in the proximity of the client, decided by the client rather than the owner
• Replication and caching lead to consistency problems as multiple copies of a resource need to be kept in sync
• Strong consistency guarantees are extremely hard or even impossible to implement in a scalable way
• Size scalability is the least problematic from a technical point of view, geographical scalability is more difficult, and administrative scalability is the most difficult