Group:
communication
Group:
systems
Topic:
hypertext as a global database
Topic:
identifying the user interface with the system
Topic:
heterogeneous database
Topic:
managing shared memory
Topic:
open systems
Topic:
replicated data
Topic:
specification and design of distributed systems
Topic:
system integration
Topic:
unique names
| |
Summary
Users want to interact with distributed systems as if it was one system. This unification can be done at file level (Vice), name level (Clearinghouse), information level, or function level. (cbb 12/88)
Subtopic: single system image 
| Quote: a single system image insulates users from distributed systems [»giffDK12_1985]
| | Quote: there should be a single UserInterface for all computing tasks [»mccrDL10_1984]
| | Quote: extraordinary ability to unify things; e.g., clutch, gear, accelerator, brake [»nelsTH_1974]
| | Quote: ZOG is a UserInterface to all computing resources [»robeG5_1981]
| | Quote: even when distributed, all services should be as if from a centralized data bank [»nelsTH_1987]
| | Quote: emphasize integration of tools and users [»akscRM5_1984]
|
Subtopic: shared state
| Quote: the primary challenge of distributed systems is shared state; shared memory is the most natural model [»cherDR3_1988]
| | Quote: make communication look like memory via ports (a queue of buffers), message screening or filtering, and logical addressing [»cherDR5_1986]
|
Subtopic: system-wide names
| Quote: a unique, system-wide name for an object provides world-wide, transparent access and simple directories
|
Subtopic: one religion
| Quote: should there be multiple religions (i.e., views of a system) or only one?
| | Quote: reality is closer to one true religion because there exists a global reality that subsumes all local views [»wegnP10_1986]
|
Subtopic: examples of unified, distributed systems
| Quote: Vice removes worries about replication of data; just store it in Vice [»boreN2_1988]
| | Quote: in Vice, mail recipients can be on any machine [»boreN2_1988]
| | Quote: Grapevine treated as one, large computer; propagation delays can surprise [»schrMD2_1984]
| | Quote: all Grapevine servers can accept and send mail from/to anyone; but delays differ [»schrMD2_1984]
| | Quote: Clearinghouse assumes one global database that is decentralized [»oppeDC7_1983]
| | Quote: clients may use any name known anywhere in Clearinghouse [»oppeDC7_1983]
| | Quote: uniform access to Clearinghouse by a stub on each client that knows a server [»oppeDC7_1983]
| | Quote: Clearinghouse easily adds new networks, adds servers, or merges networks [»oppeDC10_1981]
| | Quote: Telesophy provides location transparency; remote data same as local [»schaBR11_1987]
| | Quote: Telesophy transparently stores and retrieves knowledge [»schaBR11_1987]
| | Quote: with V, users can run interactive editors on any node without performance problems [»cherDR3_1988]
| | Quote: Amoeba gives users illusion of a single timesharing system even though it is implemented on hundreds of processors [»taneAS12_1990]
| | Quote: using wide-area network gateways, Amoeba forms a single, uniform system across possibly different countries
| | Quote: Amoeba implements remote procedure calls on each processor and in UNIX; allows transparent use of Amoeba [»vanrR10_1988]
| | Quote: a name service client sees a single hierarchy even though the database is distributed and replicated
|
Subtopic: heterogeneous database
| Quote: a heterogeneous database may need a common view despite different definitions, algorithms, and units; users can specify the mappings [»ahmeR12_1991]
| | Quote: a Pegasus schema makes a local data source appear as Pegasus database; maps between data models and query languages [»ahmeR12_1991]
| | Quote: Pegasus is a heterogeneous, multidatabase management system developed by Hewlett-Packard; object-oriented and relational [»ahmeR12_1991]
|
Subtopic: problems
Quote: Amoeba creates a file and its directory entry in separate operations; garbage collects if inaccessible (may be impractical for a huge system)
|
Related Topics
Group: communication (27 topics, 819 quotes)
Group: systems (17 topics, 530 quotes)
Topic: hypertext as a global database (30 items)
Topic: identifying the user interface with the system (16 items)
Topic: heterogeneous database (6 items)
Topic: managing shared memory (74 items)
Topic: open systems (33 items)
Topic: replicated data (51 items)
Topic: specification and design of distributed systems (14 items)
Topic: system integration (5 items)
Topic: unique names
(58 items)
|
