Skip to Content.
Sympa Menu

perfsonar-user - Re: [perfsonar-user] System/responsibility stratification.

Subject: perfSONAR User Q&A and Other Discussion

List archive

Re: [perfsonar-user] System/responsibility stratification.


Chronological Thread 
  • From: Daniel Doyle <>
  • To:
  • Cc: "" <>
  • Subject: Re: [perfsonar-user] System/responsibility stratification.
  • Date: Thu, 9 Jul 2015 09:19:03 -0400

Michael,

I'm not really sure what you mean by a "gateway level PS system" vs "targeted
PS system", so clarifying that might be a useful step. In advance though here
are some thoughts:

1.) Typically when people are setting up a mesh configuration, it's a
"everything test to everything" mode, ie a full mesh. There is also a less
common mode for meshes called "star" where a central node tests to all of the
other members but they don't test to eachother. In that respect you might be
able to generate a mesh config where the "tentacles" are testing correctly,
depending on what you're trying to accomplish. Of course you can also just
define a whole lot of meshes, even just between two hosts, as a means of
centrally controlling the regular testing assuming that each host is pulling
down the central mesh definitions.

2.) It sounds like what you are talking about here is a central measurement
archive where tests between different things report back to a MA running on a
host that may not have even been involved in the tests. This is completely
feasible and I know at least several groups are doing things like this today.
The trick is in editing the regular testing configuration so that the MA it
reports back to is pointed to whatever one you want. If that requires more
explanation I can help with that too.

Hope that helps.

-Dan

On Jul 8, 2015, at 12:12 PM, Michael Tiernan
<>
wrote:

> With the ongoing development and maturity of the PS tools, has the idea
> of stratification been discussed?
>
> Objective/Fer-instance:
> If I/MIT implements a gateway level PS system(s) properly ensconced in
> the grand unified mesh configuration and I implement my own targeted PS
> system that only tests certain things.
>
> 1) Does that targeted system get sucked up into the mesh or can that be
> prevented?
>
> 2) Can the targeted "tentacles" report back to the master level system
> to help provide a larger scale view of the involved network?
>
> Thanks for everyone's time.

Dan Doyle
GlobalNOC Software Developer
1-812-856-3892




Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail




Archive powered by MHonArc 2.6.16.

Top of Page