Net2WG/Notes/20080620

From TinyOS Wiki
Revision as of 10:33, 1 August 2012 by Gnawali (talk | contribs) (New page: ---------- = Net2 meeting 06/20/2008 = ---------- = Agenda = ---------- * TEP 4 and 135 * CTP dead files and bugs * Release = Participants = ---------- * Razvan, JHU * Om, USC * Ph...)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Net2 meeting 06/20/2008


Agenda


* TEP 4 and 135
* CTP dead files and bugs
* Release

Participants


* Razvan, JHU
* Om, USC
* Phil, Stanford
* Andreas, JHU

Action items


* TEP 4 and 135 updates (Om)
* Collection simulation tutorial (Phil)

Discussion Notes


TEP 4 and 135

Phil: The introduction of TEP135 needs more text and reference to TEP4. References are inconsistent. Section headings are off. TEP4 citations are inconsistent. Also need to refer to the t-frame i-frame TEP. Change the reserves pool: 0x70-0x75 are actually reserved

Phil: For TEP 4 in Section 3 it would be useful to say that the chair would solicit people to declare the AM IDs they need "if possible". TEP 4 should reference TEP 135 as the list of AM IDs for release 2.1

Phil: Is there a TYMO AM ID?

Om: There was but I deleted it because it is not in the release. This means that the list of applications must change (TestTymo is out)

Om: TEP4 will go through review process. However, should TEP 135 be final?

Phil: There is no steering committee which makes it hard. We should tag the current version as draft and send to the tinyos-devel for comments.

CTP dead files and bugs

Om: CtpSenderC/P.nc are dead files.

Phil: The question is what happens if one wants to have CTP *and* MultihopLQI?

Om: In this case, they are not dead code but we need to make sure they work because they have not been tested in a while.

Om: CtpSenderP.nc claims to use CtpSenderC.nc which does not exist.

Phil: CollectionSender.c works around this. But there is still a problem. The answer is that it is dead code but should be fixed. Should be fixed *after* the release.

Om: Does it make sense to have the CtpInfo.getNeighborLink/Route calls?

Phil: This is necessary for any application where it is necessary to report the topology to the gateway

Om: The argument is the index to the table of neighbors. Is this the right interface?

Phil: It is. You call getnumneighbornum and iterate through them

Om: Is your intention to return ETX values? In this can case we need to change to 'regular' ETX from EETX (offset of 10)

Phil: Fix it after the release.

Om: Add comments

Release

Om: I tested CTP with LPL, but the standard interface does not exist. Status is that CTP with LPL has been tested (will be included in the 2.1.1 release)

Om: We can include some CTP debugging scripts in the distribution. DeliveryRatio scripts makes sense

Phil: Where are the scripts going to go? TestNetwork? Sure

Om: Include it in the net2 wiki and provide information in README.

Om: I don't know where they will go but they are difficult to use due to the lack of timestamp

Om: Provide timestaps on serial

Phil: Not for now

Phil will write a tutorial on collection simulation

Om: No docs pending

Razvan: Tested Deluge with DIP and LPL. Slow but it works.