[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Actors and the Specification
Why do they use the term Actor in the CTS specification? Does this mandate a particular architecture? Why are there facets instead of WSDL? Several comments are in this space. Please add your own comments on these thoughts. tc From: Holmberg, David G. (Fed) <david.holmberg@nist.gov>
From: Toby Considine <toby.considine@outlook.com>
Too long, but I believe accurate… The Actor model treats "Actors" as the universal primitives of concurrent digital computation. The model provides both a framework for a theoretical understanding of concurrency and a basis
for implementations of concurrent systems. Principles of the Actor model include:
The advent of massive concurrency through client-cloud computing and many-core computer architectures has increased the use of the Actor model.
In the Internet of Things (IoT), the term Actor is used as it makes no assumptions of the mechanisms or even motives internal to an Actor. An Actor is simply a system that acts. The Actor
may be a traditional computer, a cloud node, a human embodied behind a user interface, or any device in the Internet of things. The Actor model is used to provide a foundation for inconsistency robust information integration. In transactive energy, we see the diversity supported by the Actor model. An energy seller may be a generator or a solar panel or a virtual power plant or a demand responsive facility or a
financial entity. An energy buyer may be home or a commercial facility or an embedded device or a microgrid or a an energy district. A marketplace acts to match tenders, but may also buy or sell power. An energy storage system may act as a buyer or as a seller
at any time. The common transactive services are intended to be common to all actors in Resource markets. We use the term “facet” to name a set of interactions that an Actor may use in a resource market. An actor submits tenders to buy or to sell. An Actor may operate a market. An Actor may provide
telemetry on metering; that telemetry may be collocated with the Actor buying or selling, or it may be on another Actor, perhaps owned by another party, that is registered as associated with the market participating Actor. This specification makes no requirement
as to how these facets There is a growing use of the term “cloud-native computing” to name extending the architecture and technologies developed for use in central clouds beyond data centers to remote systems, including
to IoT devices. As best practices in the cloud evolve to zero-trust within distributed applications, the distinctions of within or without the cloud become less important. A discussion of the rapidly-evolving topic of cloud-native computing is beyond the scope
of this specification. This specification does not require that implementations conform to cloud-native computing. The authors of this specification have been informed by cloud-native computing, and we use the term “facet” to indicate message end-points that
an implementer may distribute over one or more Actors to build a conforming application.
From: William Cox <wtcox@coxsoftwarearchitects.com>
BTW the paper you sent earlier was from the Association for Computing Machinery (ACM) Design Automation Conference 1999, not an IEEE event, and is at a useful level. Reference (in the paper) is
ACM Reference Format:
Note the narrow focus of the actor history on realtime systems, which I think is a defect.
The IEEE draft is not IMO very compelling - I don't think it's easily referenceable, and it's convoluted. given all of the metadata, in fact I think it's confusing. (Smells a bit of TOGAF as well). And its focus (when it gets to a reviewable state) is describing
multidimensional spaces of actors. References are largely useless for actor system practitioners.
Of the two, the ACM DAC paper is far superior as a reference, as the succinct early section has decent references.
With a broader search, there's the original formulation by Hewitt 1973 (mis-cited in Lohstroh unless there actually is a "Standford, CA"), paper attached. I don't think this (though first) is a very useful reference. Paper attached.
I find Hewitt 2010-2015 to be a better introduction - attached. Updated to include IOT discussion and more, and a more general introduction/backgrounder than Lohstroh.
Thanks!
bill
On 1/21/22 10:55 AM, Toby Considine wrote:
I am thinking of stealing some language from this:
See sections 5.3 and 5.4
tc
--
William Cox
Email: wtcox@CoxSoftwareArchitects.com
Web: http://www.CoxSoftwareArchitects.com
+1 862 485 3696 mobile |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]