BioInfoBank Library


LectureResultsQuestionsForumPDF
 

Architecture and characterization of distributed systems

Disciplines: Distributed computing
Introduction:

lech

The lecture will be architecture and characteristics of distributed systems[1][21]. A distributed control system is a collection of independent machines or automatic interconnected into a single, coherent, logical whole. Most of these devices are computers connected through a local area network(LAN)[2], or wide area network(WAN)[3] or metropolitan(MAN)[25]. The basic advantages of distributed systems include resource sharing[4] and transparency[5]. Transparency is characterized mainly by the fact that it gives a single user group, and their illusion of a centralized, compact, consistent operating system resistant to various types of failures.
We can distinguish the following types of transparency in a distributed system:

  1. Location transparency[6]
  2. Migration transparency[7]
  3. Relocation transparency[8]
  4. Replication transparency[9]
  5. Concurrent transparency[8]
  6. Failure transparency[10]
  7. Persistence transparency[11]
  8. Security transparency[12]

The basic features of a distributed system include:

  1. resource sharing[4], including files, printers and other peripherals
  2. openness to the extension or expansion of different manufacturers in terms of system hardware and software
  3. concurrency[13] or perform many processes and tasks simultaneously
  4. scalability[14] is the ability to expand with new computers or technical equipment
  5. fault tolerance[15], namely, a coherent system despite the failure of individual computers in the cluster[23] or the grid[24]
  6. transparency is the perception of the system as a whole by the user (hiding the heterogeneity[26] of users, operating systems, differences in communication and organization)

One of the main advantages and differences that occur in distributed systems is the presence of an middleware[16], which mediates the communication between applications.

Examples of middleware are:

  1. Socket[17]
  2. RPC (Remote Procedure Call)
  3. DCE (Distributed Computing Environment)[18]
  4. CORBA (Common Object Request Broker Architecture)
  5. DCOM (Distributed Component Object Model)[19]
  6. RMI (Remote Method Invocation)[20]

Practical exercises:

  1. Install openMosix on two computers. You can either use a permanent installation on the hard disk or from a LiveCD distribution. The combination of follow through hub or switch.
  2. Share a printer or folder in Windows to another computer on the network.

 



Available materials: The following materials and learning resources are available in this lecture:

Wikipedia material [1] Distributed computing

Wikipedia material [2] Local Area Network

Wikipedia material [3] Wide Area Network

Wikipedia material [5] Transparency

Wikipedia material [6] Location transparency

Text [7] Migration transparency

Text [8] Relocation transparency

Text [9] Replication transparency

Text [9] Replication transparency (additionally)

Wikipedia material [10] Failure transparency

Wikipedia material [11] Persistence transparency

Wikipedia material [12] Security transparency

Wikipedia material [4] Shared resources

Wikipedia material [13] Concurrency

Wikipedia material [14] Scalability

Wikipedia material [15] Fault tolerance

Wikipedia material [16] Middleware

Wikipedia material [17] Socket

Wikipedia material [18] Remote Procedure Call

Wikipedia material [19] Distributed Computing Environment

Wikipedia material [20] Common Object Request Broker Architecture

Wikipedia material [21] Distributed Component Object Model

Wikipedia material [22] Java Remote Method Invocation

Wikipedia material [23] Cluster

Wikipedia material [24] Grid

Wikipedia material [25] MAN

Wikipedia material [26] Heterogeneity

Please read carefully every provided learning resource and material. And after, if you feel self confident enough, you can:
Test your knowledge:
Or view solutions: Show all questions and answers!


2013-06-19 11:56:32 © BioInfoBank Institute