Page tree
Skip to end of metadata
Go to start of metadata

Network diagram:

Source on draw.io here.

Spreadsheet with IPs and hostnames for the DSMs and ubiquitis is here.

Files like /etc/hosts tables and the ansible hosts are generated from the cheesehead.xml and hosts.xml files in the project config directory, projects/CHEESEHEAD/ISFS/config.

Network naming key:

Name PatternDevice category
<site>bDSMs at the bottom of the tower, all Titans except for one Viper.
<site>t

DSMs at the top of the tower, all Raspberry Pi.

u<site>Ubiquiti radios serving as clients, aka station adaptors (SA).
tv[123]Three Ubiquiti radios on the TV tower serving as access points (AP).
<site>DSMs at sites with only a single DSM, all Raspberry Pi.
base,iss,...Ubiquiti SA radios for ISFS base trailer, ISS, and other guest platforms.

Accessing ustar:

The ssh config below should work to connect to ustar during staging at EOL, and possibly during operations depending on what ports we can forward on the public-facing router.

Host ustar
User daq
Hostname ustar-isfs.dyndns.org
Port 23
HostKeyAlias ustar
CheckHostIP no
StrictHostKeyChecking yes
LocalForward 15380 localhost:80

The ssh tunnel is also up and working, so this configuration should work both in staging and during operations.

Host ustar-tunnel
User daq
Hostname eol-rt-data.eol.ucar.edu
Port 31022
HostKeyAlias ustar
CheckHostIP no
StrictHostKeyChecking yes
LocalForward 15380 localhost:80

From ustar it is possible to ssh to any of the DSMs by name.

The LocalForward allows the local browser to connect to Nagios on ustar using this URL: http://localhost:15380/nagios/.   Use nagiosadmin and sun-dog.

Aiming nanobeams:

Station connection status:

StationConnection methodWorking?NotesRadiosUplink APPower SchedulePlans
NW1ubiquitiyesinstalled G2 relay AP for nw2/3, but neither have been able to connect. unw1ap radio has been removed and unw1 connected directly to dsm.G2 unw1tv

NW2cellyes40' Rohn, cell signal at the base of the tower (verizon).  Could not connect to relays at either nw1 or nw4.  Installed Verizon cell 7/26.G2 unw2, turned off 7/26

Add a directional cell antenna when available.

NW3cellyesPAM. USB modem with AT&T SIM, directional antenna. Web interface reports 3 bars, 4G LTE.

AT&T Cell

G2 unw3



Did this radio already got removed?
NW4ubiquitiyesPoE passthru from unw4 to unw4ap turned off 7/27

G2 unw4

GU unw4ap, off

tv

NE1ubiquitiyeswinch repaired, raised, une1 connects to tvG2 une1tv

NE2ubiquitiyes
G2 une2tv

NE3ubiquitiyes

relay installed: G2 une3ap as AP for une4

G2 une3
G2 une3ap
tvoff 9-5
NE4ubiquitiyestoo much high ground between site and WLEF, use une3ap relayG2 une4une3apoff 9-5
SE2ubiquitiyes
G2 use2tv

SE3ubiquitiyesG2 use3ap installed bearing 125 degrees to catch se6 and also se4, but se4 signal is weak; the relay to both se6 and se4 might be better with a LiteAP AC, but I didn't order oneG2 use3
G2 use3ap
tvoff 9-5
SE4ubiquitiyesinstalled G2 use4, has weak signal to use3apG2 use4

use3ap

off 9-5
SE5ubiquitiyes40' Rohn, Gordon climbed to adjust bearing to 160 degrees, sufficient signal to use6ap relayG2 use5use6apoff 9-5
SE6ubiquitiyeslocked to AP use3ap; turning off power overnightG2 use6
G2 use6ap
use3apoff 9-5
SW1cellyes

ISFS-M21 166.167.71.187, sw1-isfs.dyndns.org

radio removed; rsync working from ustar over cell (ISFS-303)





SW2ubiquitiyes100' Rohn, sees both tv1 and tv2, apparently just needed to be switched to PtMP modeG2 usw2tv

SW3ubiquitiyesrelay radio removed

G2 usw3

tv

SW4ubiquitiyes
G2 usw4tv

baseubiquitiyes
G2 basetv

Radio model key:

  • G2 - NanoBeam AC Gen2, model NBE-5AC-GEN2-US
  • M16 - NanoBeam M5, model NBE-M5-16
  • M5 - NBE-5AC-16

Power Conservation

The current thinking is to turn off radios overnight on towers with two radios to reduce power consumption, and then also at sites which depend on those relays.  For example, turning off the relay at se3 cuts off three sites: se3, se6, and s5.  See Radio power schedule installed blog entry.

Network Plan

The final sites have been added to the net using cell modems: nw2 and nw3, Rohn and PAM sites with a single DSM at the ground.  The USB sticks at those sites are easy to access and swap.  Gordon tried an AP at nw4 to see if it catches nw2 and nw3, but that didn't work either.  At the moment, nw3 has a directional antenna.  It would probably help nw2 to have a directional antenna also, once one is shipped to the field.

se5 is online, but barely.  The rsync seems to be keeping up mostly, but sometimes it can fall behind by several hours.

If data from a tower trailer needs to be collected by sneaker net, then we have to come up with a scheme like running an rsync on site (very slow) or continuously sync'ing data from the top DSM to the bottom DSM and then periodically swapping the bottom USB stick. At the moment that seems unlikely to be needed.


  • No labels