perfsonar development work

["Jeff W. Boote" <>]

Roman Lapacz wrote:
> Nicolas Simar wrote:
> > Hi,
> >
> > In the RRD MA configuration file, the person who build the metadata 
> > configuration file  can chose the "Data storage unit".
> >
> > The visualisation tools querry the information from the MA.
> >
> > What happens if one RRD MA instance export bps, another one kBps?
> > 1) Do the visualisation tool have to deal with it?
> > 2) Do the MA have to normalise it?
> > 3) Are there best practices we should suggest to ensure consistency in 
> > the service offered globally?
>
>
> I think we should have one terminology for units to avoid the situation 
> when the same unit will be called differently in two instances of RRD MA 
> (example bps and bit/s).

I agree. See that attached message for suggestions on where we come up with unit 
names from. (From the last time this topic came up.)

> Transformation between units should be done by transformation service 
> (TrS). No need to add this functionality to MA (it 's possible but may 
> increase the time of responses from MA).

As I mentioned in my other message. The other possibility is to have the client 
do it.

jeff

--- Begin Message ---

• From: "Jeff W. Boote" <>
• To: Vedrin Jeliazkov <>
• Cc: Loukik Kudarimoti <>, Roman Lapacz <>,
• Subject: Re: [pS-dev] [Fwd: units expected from RRD MA]
• Date: Fri, 11 Aug 2006 10:16:42 -0600

Vedrin Jeliazkov wrote:

Hi,

Nina and myself agree that:

- 'unit' is too generic and calling it <*>valueUnit (as suggested by Jeff)
would be better;

- it's preferrable to have <*>valueUnit in datum (as suggested by Loukik),
rather than in metadata;

That is fine. It just seems wasteful to repeat a field like that in every single datum when it is common across the full set of data. As I said before, I don't feel strongly about it.

We also feel that opening the door for free text content of <*>valueUnit could
quickly lead to a multitude of variants for a given unit name in different
service instances. It would be much better if the allowed units are well
defined, known by clients and enforced by respective services in some way. We
would very much like to avoid the possibility to have simultaneously:

b/s
bit/sec
bits/sec
bits per second
etc...

I agree. For many of the units there are ISO (or at least SI) standards in place. (For time units for example.) We should use them when we can. And for the others we should make very sure that we have looked. One good place to start is:

http://en.wikipedia.org/wiki/SI
http://en.wikipedia.org/wiki/Binary_prefix

And, as much as we might think it is confusing. The most common specification of bytes vs bits out there is B vs b. We should stay consistent with the rest of the world here.

And last, but not least, maybe it would be better to have bytes per second
(instead of bits per second) in the default rrd-database_TEST.xml, because
data sources reporting bytes seem to be more common.

We should stay consistent with the default units of the SNMP mib value - which I believe is octets (bytes) in this case.

In general, I don't think we should do units conversion unless it is somehow requested. (Laziness is a virtue in programming. ;) )

jeff


--- End Message ---