Pipelining
From Erights
(Difference between revisions)
Kevin Reid (Talk | contribs) (briefly describe pipelining -- but there's surely a document about it already?) |
|||
| Line 5: | Line 5: | ||
Note that this applies to the promises resulting from [[eventual send|eventual sends]]; for example, in <var>a</var> ← <code>b</code>() ← <code>c</code>(), the <code>c</code>() message may be transferred to <var>a</var>'s [[vat]] before <code>b</code>() has been delivered to <var>a</var>. | Note that this applies to the promises resulting from [[eventual send|eventual sends]]; for example, in <var>a</var> ← <code>b</code>() ← <code>c</code>(), the <code>c</code>() message may be transferred to <var>a</var>'s [[vat]] before <code>b</code>() has been delivered to <var>a</var>. | ||
| - | + | ''XXX There must be a page on www.erights.org about this — find it and link to it.'' | |
Revision as of 19:40, 18 April 2007
Pipelining is the property that messages sent to remote promises are transferred to the vat currently holding the resolver of the promise, as opposed to being buffered until the promise is resolved.
This allows messages to be delivered to the resolution of the promise with fewer network delays.
Note that this applies to the promises resulting from eventual sends; for example, in a ← b() ← c(), the c() message may be transferred to a's vat before b() has been delivered to a.
XXX There must be a page on www.erights.org about this — find it and link to it.
