Glen,
I'm struggling with a number of statements, I agree the connection between
the user and the cloud is one part of the model, and the other part is the
connection of the cloud to a logical print something or other (within my
system). Since the model started as an imaging model, I believe the cloud
to print connection should also be usable for a scan to cloud, with
different elements.
As a user and implementer I'm focusing on where the connection to the cloud
originates, and where the connection from the cloud is terminates.
So where you have in the last drawing - Vendor M and Vendor N, are those
print services from different print vendors or?
Larry
From: cloud-bounces at pwg.org [mailto:cloud-bounces at pwg.org] On Behalf Of
Petrie, Glen
Sent: Friday, April 06, 2012 10:35 AM
To: Michael Sweet
Cc: ipp at pwg.org; cloud at pwg.org
Subject: [Cloud] RE: [IPP] Where is the Cloud Print Manager in the PWG:
Cloud Model
Mike,
Thanks for the update. I was reviewing the diagram again and I had noted
there was no actual print entity "in the cloud"; that is, as an SAAS in
cloud terminology; so adding the Cloud Print Service at least provides a
cloud print SAAS.
Some more discussion ----------
I am still having a problem with the use of a Cloud Print Manager (in the
printer) concept (an the name) in the general context of being a cloud
citizen and with associating a User with an individual Print Service
(Printers). That is, it is not modeled from User's, use-case or Cloud
Provider developers' point of view. As currently defined, the current Cloud
Print Manager is an artifact of a particular implementation (like DocuPrint;
maybe even CUPS) for a fan out with a single controller (Print Servcie?);
from a reference model or architecture point of view it should not matter if
there is such an entity providing this function. This means your simple
diagram reduces to
Cloud Print Provider
Client <---> Cloud Print Service <---> Printer
With """"""maybe""""" the Printer = [ Cloud Print Manager <---> Printer(s) ]
as seen from Cloud Print Provider
Actually, I see no reason to add the Print qualifier to the Provider; that
is,
Cloud Provider
Client <---> Cloud Print Service <---> Printer
This model states that a Client, as a proxy for the User, can access a Print
Service, of the User's Cloud Provider, to print to a User's register print
device.
This model can be used to represent several different solutions or
implementations:
1. The "Cloud Provider" is Print Cloud (like HP's or Epson's or)
Print Cloud Provider
Client <---> Cloud Print Service <---> Printer
a. A good and bad model - requires other Cloud Solution to federate
with Print Cloud Provider. This would be a big development effort for
print vendors.
2. Cloud Print Services are integrated into a Cloud Provider like
Google Cloud
Cloud Provider
Client <---> Cloud Print Service <---> Printer
a. Again, good and bad. Good for print vendor in that the Cloud
Provider manages Print Service; bad for the same reason.
3. The Cloud Print Service is a "mini-cloud" or a "plug-in cloud" (but
still in the Cloud) that is callable by a Cloud Provider.
Cloud Provider
Client <---> \ \
\ \
\ Cloud Print Service <---> Printer
Where there is cloud around both Cloud Print and Cloud Print Service
a. Yes, this looks like model in the current specification but with
some slight differences.
i. The
Cloud Print Services are in the cloud; thus, they are cloud SAAS
1. Yes, the Print Service could be in the physical printer but
virtually linked ("registered") in the Cloud
ii. A
printer is associated with a single Cloud Print Service
1. >From the User's and the Cloud Provider's point-of-view, one Print
Service is associated with one Printer. The User and Cloud Provider
developers never know about fan-out, fin-in, fan-up, fan-down or any kind of
fan; they know there an associated Print Services for a specific Printer.
a. Internally, an individual "Print Service" can support 2, 700,039
Printers; that is something the print vendor does as part of their
implementation detail. It will not affect the API's the Cloud Provider
developers uses to print to the printer. And has nothing to do with the PWG
Cloud Print specification.
2. >From the User's and the Cloud Provider's point-of-view, one Print
Service (associated with one Printer) has its own queue.
a. If a print vendor wants to have individual queues (the "bucket" of
Print Jobs) for each printer or have one giant queue (a big "bucket") that
individual printers pull jobs from, this again, is a print vendor
implementation detail. And has nothing to do with the PWG Cloud Print
specification.
iii. From
the Cloud Provider's point-of-view, the above definitions makes all Cloud
Print Service interfaces look the same; that it,
1. An individual Printer is registered and has an associated a single
Print Service
2. The Cloud Provider only talks to the Printer's Print Service for all
APIs and actions.
3. Print Jobs are submitted to the Print Service for the target printer
(no matter how the queue is implemented)
iv. Thus,
the goal of the PWG Cloud activity is to define the API's and data for the
Cloud Provider to/from a Cloud Print Service
1. I know - use IPP
a. I don't think IPP will be used by everyone. (Who is using it now
for Cloud?) What "will be used" is the PWG:PJT, which does not require IPP
to be useful. PWG:PJT make federation of the Print Jobs very simple. In
general PWG:PJT is a perfect candidate for cloud print.
So the final model is something like
Cloud Provider |------- a
Directory/Registry of Vendor Print Clouds
Client <---> \ \ < -------------------- register
Printers (and its Print Service) to a User
\ \
\ Vendor M Print Cloud
\ Vendor N Print Cloud
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
Print Service <---> Printer
glen
_____
From: Michael Sweet [mailto:msweet at apple.com]
Sent: Friday, April 06, 2012 9:00 AM
To: Petrie, Glen
Cc: Peter Zehler; Ira Mcdonald; ipp at pwg.org; cloud at pwg.org
Subject: Re: [IPP] Where is the Cloud Print Manager in the PWG: Cloud Model
[Adding cloud at pwg.org since this is a cloud discussion]
Glen,
In the original BOF-generated functional model, the Cloud Print Manager is
either located in the Printer or attached to the Printer (e.g. in the local
print server). The Cloud Print Manager manages communications between the
lower-level printer interface and the Cloud Print Provider.
Based on our last telecon, we have introduced a Cloud Print Service into the
model which is the service in/under the Cloud Print Provider (which acts as
a System Control Service in the Semantic Model sense) that manages the
jobs/queue in the cloud for the Cloud Print Manager. The rough ASCII art
diagram in my mind looks like this:
Cloud Print Provider
Client <---> Cloud Print Service <---> Cloud Print Manager <--->
Printer(s)
Thus, Clients and Cloud Print Managers talk directly to the Cloud Print
Provider and corresponding Cloud Print Service, but never directly to each
other, and only the Cloud Print Manager actually talks directly with the
Printer(s) (physical or logical) it is registering/sharing with the Cloud
Print Provider/Service.
There can potentially be multiple logical or physical printers behind the
Cloud Print Manager (think traditional fan-out configurations and
reprographic services) and those printers *may* be addressable using the
IPP/SM output device attributes/elements, but there is only a single Cloud
Print Service per Cloud Print Manager and to the Client it will appear to be
a single "queue" with one or more output devices associated with it.
On Apr 6, 2012, at 8:34 AM, "Petrie, Glen" <glen.petrie at eitc.epson.com>
wrote:
Ira, Pete,
I believe it was stated in the last conference call that the Cloud Print
Manager in the cloud model diagram "is not in the cloud". Is this true?
If so, where is it "located"?
Glen
--
This message has been scanned for viruses and
dangerous content by <http://www.mailscanner.info/> MailScanner, and is
believed to be clean. _______________________________________________
ipp mailing list
<mailto:ipp at pwg.org> ipp at pwg.org
<https://www.pwg.org/mailman/listinfo/ipp>
https://www.pwg.org/mailman/listinfo/ipp
__________________________________________________
Michael Sweet, Senior Printing System Engineer, PWG Chair
--
This message has been scanned for viruses and
dangerous content by <http://www.mailscanner.info/> MailScanner, and is
believed to be clean.
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pwg.org/pipermail/ipp/attachments/20120406/8a34e75d/attachment-0001.html>