RFC 530 UNKNOWN MONITORING · INFRASTRUCTURE
Report on the Survey Project
Regular monitoring of ARPANET host availability
RFC EDITOR
Bhushan, A.K., "Report on the Survey Project", RFC 530, DOI 10.17487/RFC0530, June 1973, <https://www.rfc-editor.org/info/rfc530>.
BIBTEX
@misc{rfc530,
series = {Request for Comments},
number = 530,
howpublished = {RFC 530},
publisher = {RFC Editor},
doi = {10.17487/RFC0530},
url = {https://www.rfc-editor.org/info/rfc530},
author = {A.K. Bhushan},
title = {{Report on the Survey Project}},
year = 1973,
month = jun,
} RFC text
Contents
This text was converted from the RFC Editor's PDF edition and may contain conversion errors; the layout is approximate. The PDF is the reference copy.
.. -
Received at NIC July 3, 1973 1737 5
~
JU L 1 1973
Ne tv.o r k \!orking Group Abhay 8hushan
Re qu e s t Fa r Comr.ie n t s # 530 t·i1 T Proj ec t r · ~AC
HIC # 17375 22 June 1973
Updates: RFC 308, 523
A REPORT Or! TilE SURVEY PROJECT postel
The purpose of this paper is 1) · to report on the status of
the SliHVEY project and current data, 2) to l n f o rrn the ARPAi~ET
cor, lr,lUnityof the services VJe offer related to this project, 3) to
report on our future plans, and 4) to ask for suggestions and
Imp rov er.ie n t s •
I. THE STATUS OF TI!E SURVEY PRO GRM·j
The SURVEY program is in operation again as man v of you r.iav
have noticed. \:e are surveying at 20-minute intervals nU\1 (vie
were surveying at 15-~inute intervals before). The SUnV EY
p r o g r ar.i a t t ernp t s to do a cornp l e t e rep Cl n l t l o l connection
protocol) to the LOGGER socket (socket 1) of each host 1 isted in
the survey table. lie currently survey all the known "SERVER"
I to s t s on the i\RPAilET. It is easy for us to add and delete hosts
f r om this table.
The previous version of the SURVEY prograr:1 ~as abortin~ in
the m l d d l e of lep, but nov. we a t t emp t to complete ICP (v. l t h a 5
second time-out ror each of the connections after receiving the
server socket number "S" on the ICP connection). This latter
change wa s r.iad e as rnan y II0STs do no t p r ov i de for either time-outs
v
or queu i ng of the ilCP c omrnan d s (RFC IS). Consequen t 1 y,
l n cor.ip l e t e ICPs v.e r e tying up HOST connection resources. Please
let me know i f our 5-second t l rne-iou t is too srna l l and i f you are
encountering p r o b l er.rs "lith SURVEY.
The survey p r o g r arn records date, t l me , status, and response
t lme for each host (the re s pon s e t lr .ie is currently averaged, and
da t e and t lr.ie are only recorded for the .f i r s t survey in a block
of ~G surveys). \Ie are in the process of modifying the SCRVEY
recordin6 pr:ocedures to record all the data for each inuiviJual
survey test in a stanJarJ ~SCI I format suitable for transmission
to the [j,"T :\COt·1PUTER (CCA) vrhe r e vie are storing the SllRVEY data.
This database at the CCA-D ATACOI·:PUTER is accessible via p r o g r arns
in our !;UDDLE subsystem. A future memo vii 11 e x p l a in the use of
the DATACOI·1PLTER database and faci 1 i t ies to access it.
The response t l rne measurement is the tine taken to e s t ab l ish
the full-duplex TELf ;ET connections. It is not the tine taken to
r; e t the pro••• o t r.ie s s a g e , or lo r;gin ~~ in. The latter rie a s u r e s of
response t lrue are r.io r e raean i n g f u l to the u s e r but our rnea s u r e for
r e s non s o t lr.ie represents a l owe r bound for tile ' "true" response
PAGE 2
t l r.ie • It should be noted that our r e s pon s c t l me rnea s u r l n g
procedure h a s a l t e r e d slightly (prev iously vie v.e r e me a s u r i n g the
t l r.ie taken to get the server socket n urnb c r "S" on the lep
connection>. It should also be noted that the response ti me
me a s u r e r.ie n t is val iJ only when the I!ost state is "logzer
available".
J I. r,
I,! ET \ J0 K STAT USSE RV ICE 0 t J SOC l( ET 1:;
\I e a net \"10 r k s tat u sse r vic eon soc k e t 15 ( dec i n: a 1 > for
0 f fer
results of the "last'! .o r t he no s t recent
comrnun i c a t i n g the
survey. To use this service, just l Cf' to HOST 70 socket 15.
The SLRGIV p r c g r art v. l l l t r an sr.i i t data in the f o l l ow i rig f o rma t :
mo nth, day, yea r , h 0 u r s , r;l j 11 u t e s
host,status,response-ti~e
ho s t s s t a t us , r e s pon s e t l me
.. . .. .. . .
r
.'
-1
An e x ar.ip l e of the data t r an sm l t t e d is given b e l ow ;
ilconnection to host dr.icg socket 15
cor.ip 1 e ted.
OG,J5,1973,14,44
001 ·,5,013
002,5,031
003,1,200
004,5,015
UUG,5;011
007,3,20U
008,1,2JO
009,5,002
010,3,200
011,S,U03
013,5,01G
014,5,002
015,5,014
OiG,5,u;;1
023,3,011
031,1,2JO
032,5,U14
035,5,015
UG5,5,uJ8
OGG,5,J1u
Ull9,5,J2u
07u,5,J09
u74,5,u11
07/),5,002
08G,3,2JU
PAGE 3
133,1,200
134,5,QUG
138,4,200
198,5,007
-1
COflllECT 10[1$ ABORTED
CLOSED [3 Y F0 r\ E I GI~ H0 ST
The l iost number is in de c l ma l , the response t l rie is In tenths
or seconJs, and tile status co de isas described be l ow :
Code 5 - IC? to socket 1 successfullY completed,
logger availa~le.
'Co d e 4 - ICP t i ru e d-r ou t (20 ' s e con d s L,
lo~ger not res ponding.
Code 3 - ICP aborted by foreign l.o s t ,
logger rejecting.
Code 2 - flo reset response f r or .i f o r e izn lio s t ,
(15 sec tir.lc-out) ncp not rcspondin [;.
Code 1 - Fo r e l g n host DEAD status returned by IflP
Host dead or disconnected.
Code 0 - Ilo t surveyed or' un de t e rra l ne d ,
All t ran s m iss ion i sin s tan dar J 8 - bit net VJO r k AS C I I b y t e s •
-- A response tir.le of 200 (20 seconds) is sent for all other status
except 5 (logger available). The entire survey data is for the
time recorded on the first 1 ine (the survey usually takes only a
co u p 1 e 0 f r, r l nut e s ), and s h 0 u 1 d beg 0 0 d to VJ i t h i n 20m i nut e s (1 0
ml nu t e s average).
The above service is designeJ ~ore ·for use by programs but
can also be used directly by human users. Other hosts on the
A HP1\fJ [T r,I a y per i 0 J i c a l l y colle c t the sur v e yin f 0 rr .1a t ion fro m u s
anJ store and display the l n f o rma t l on as t he y please. They can
also Jisplay to their user's the latest host availability
l n f o rr.ia t l on w i t hou t actually doing the survey t her.ise l v e s ,
III. ACCE:';SIIIG SUI\VEY D/\T'\ Vii' TilE Ir:ET\:RK" PROGRAf·l
To use the I;ET~:RI~ p r og r am connect to (via fCP) socket 1 of
host 70 (i.e., login), and login by typing "I o a t n <host
no>initials CI\." (e • .~., login 70akb <CR)>. After you are l og g e d
in and receive the ";" o r or.ip t , invoke f!ET\Ir.K by typing "r 'ETl;RK
<cn>" «Cr<> = Carriage Return). You Vii 11 get the ri e s s axe
"Ile two r k cor.iman d s available, and the "Q" p r omo t f r ort CALICO
fJETlmK. i·JoYJ type any of tile f o l l ov.Ln g comr.ian d s and expect to
receive t ype ou t of tile f o rm shown be l ov. :
r
(comments are in parenthesis)
PAGE 4
1. OISPLJ\Ylr~G TilE RESULTS OF TilE "L/\ST l I . S L R V , I : l
QSURVey (you type " s u r v <CR)II)
SUHVEY TAKEIJ AT 14:44:51 Ot~ 06/05/73
---HOST--- ---rr--- ----STATUS-----
n:
OCT DEC
uc 1 a -umc 1 001 Logger available
sri-arc 2 002 Logger available
ucs b-r.to d Zf 003 003 l .o s t disconnected
u.tah-10 I~ uO 4 Logger available
I.l i t -rnu 1tic DOG OOG Logger available
rand-rcc 7 007 Lo g ge r not r e s pon d i n g
suc-auept ILl G08 l .o s t disconnected
harv-10 11 009 Lor,her available
11 - G7 12 01U Logger not responding
su-ai 13 all Logger available
case-10 15 013 Lo~ger available
Cf,1U-OCC 16 014 Logger available
i 11 i ac . 17 015 Logger available
ar.ie s G7
r 20 01G Logger available
usc-44 27 023 Logger available
cca 37 031 ho s t disconnected
p ar c -ma xc 40 032 Logger available
ucsJ-cc 43 035 l.o g ge r ava i 1abl e
uc1a-ccn 101 JGS Lo~ger available
sri-ai 102 JGt) Logger avaiiable
bbn-tenexa 105 uG9 Logger available
rn i tr dmc g 10C 07J Logger available
l'-tx2 112 074 Log ~er available
emu -lOa 1 t 11G 078 Logger available
usc-isi 12G OSG Logger not responding
bbn-tenexb 205 133 Host disconnected
mit-ai 20G 134 Log~er available
l'-tsp 212 133 Logger rejecting
r.l i t -m l 30G 198 Logger available
2. DISPLAYlr :G THE CURREr:r SLl~r:;\RY
QSUl.1r.la ry. of. su rv e y s (you type sun <CR»)
14 Surveys from 10:23:37 on OC/05/73 to 14:44:51 on OC/05/73
---I-;OST--- ---J---
OCT DEC
-~~-UP- -RESP
ucl a-rime lJJl Ou1 U9 3 ~~ 01.25
sri-arc 002 002 ' 07 9 ~~ 02.37
uc s b-r.io.I 75 aU3 DO) o2 ~ ~~ 00.'67
utah-10 OU .. 004 o7 ~;~ 01.43
' ..o uuli 100 ~~
rod t vmu l tic uu G1.10
ranJ-rcc 007 uLl7 00 U ~. 00 '. I) 0
-, PAGE 5
sde-a<.lept J10 008 000 ~~ 00.80
harv-10 011 009 1 J a ~~ 00.19
ll-G7 012 U10 oor~ 1~.17
su-ai 013 all 100';.; 00.34
ease-10 015 013 100 % 01. 53
emu-ee 01G U14 o93 ~~ 00.22
i 11 i ae 017 015 U71 ~~ 01. 34
ames-G7 020 u1G 'J79~~ 02.39
u.se- 44 027 023 osnG UO.97
eea 037 031 o2 9?~ 02.15
pa r c-ruax c 040 032 1 OJ ~~ u1.3 Ll
ues<.l-ee 043 035 o71 ~~ 01.57
uela-een 101 LlGS U9 3:~ uo.35
sri-ai 102 JuG 093 :; 00.99
bbn-tenexa 105 lJG9 100% 01.92
mi t -Jmeg lOG 070 100 ~6 00.38 *
ll-tx2 112 074 07~% 01.08
emu-10alt 11G 073 08 G~~ 00.24
use-isi 12G 036 043 ;~ 11.32
bbn-tenexb 205 133 f)uO~~ OJ.OO
r.i l t - a i 20G 134 10 O~; aO.G3
ll-tsp 212 138 00 0 ~~ 00.00
r.l it -roll 30G 193 1\)0 ~; 00.G7
\.'as really up 100 ~~ as it eOr.lpleted 14 surveys
*1-11 T-[)!·iCG
at "20 ninute intervals be tvze e n 1023 and 1444.
3. D I SPU\Y I rJG TilE LOi!GTERf·1 SUfTARY
QL at I Gte nil. SUI· ir.'a r y • 0 f • sur v e ys (you type "1ong(SP)sum("SP>")
23232 Surveys from 19:43:24 on 04/27/72 to 10:03:32 on OG/OS/73
---1:05T--- ---#--- -;;-UP- -RESP
OCT OEC
ue 1a-mole 001 001 o7 5 ~~ U1.0G
s rl a r c
e 002 002 oG9;~ 01.70
uc s b-r.io d z S J03 003 05 G;~ 00.G9
utah-10 0.04 004 o71 ~~ 02.02
bbn-nee 005 005 000;; OO.OJ
rn i t -r.1U 1 tic OOG aOli OLJ5;~ 04.52
ranJ-ree ;)07 007 oUa;" 00.30
sJe-aJept 010 00.3 00 6 ~o 01.G7
harv-10 011 uu9 aG8 ~~ 00.17
ll-G7 012 u10 01 G ;~ 04.9~
su-ai 013 011 J 7 G:6 00.41
ease-10 015 013 o3 9;~ 00.75
cr.iu-ic c 01G 014 1.J7 5 :" 00.20
ill i ae 017 015 05 o:~ 02.95
PAGE G
.
ames-G7 020 01G 045 ;~ 01.51
raJc-G45 · 022 013 OOO;~ uo.aD
nbs-ccst 023 019 000 ~~ 00.J2
usc-44 027 023 0191; 00.36
cca 037 031 o3 O;~ 01.13
p a r c -rnaxc 040 032 00 1 ~~ 01.34
ucsd-cc 043 035 1115~~ 02.14 (also add results of Host 129)
ucla-ccn 101 'H.i 5 oa 1;~ 00.32
sri-ai 102 DGG 01 G~6 00.94
bbn-tenexa 105 OG9 07 3 ~6 01.13
.r,l it -dr.lcf, lOG 070 o9C ~~ 01.05 (real UP t l me is 'a bo u t 80;6)
rand-csf, 107 071 G2 3 ~~ 01.32 (later changed to usc-isi)
ll-tx2 112 074 00 0 ~~ 00.91
cmu-s I Oa 1 t 11G 078 035 ;~ 00.20
usc"':isi 12G 08G 03 a ~~ 01.00 (previously Rand-CSG)
???????? 201 129 00 8?~ 02.01 (the old UCSD-CC Host)
bbn-tenexb 205 133 o19;G 01.56
I.lit-ai 20G 13 1• 08 4;~ 00.82
l l r t so : 212 138 00 O~~ 00.08
r.i l t -rn l 30G 1S3 073::; 00.73
4. D..l.S.tlAY I r j G If IE CUR RErJ I II 1ST0 RY 0 F l\ I ;0 SI
QlilStory of l·llC (you type Ilhist<SP>nic<CR>t1)
(You (,lay use any acceptable host n arne or number)
14 Surveys taken beginning 10:23:37 on OG/OS/73
Test ";01 LOGGER available
Test ;102 LOGGER available
Test #03 LOGGER available
Test #04 LOGGER available
Test Tl05 LOGGER not responding
Test #QG lio s t dead
Test #07 lio s t dead
Test iIOS LOGLER available
Test. #09 LOGGER available
Test ilI0 LOGGER available
Test ;/11 LOGGER available
Test #12 LOGGER available
Test #13 t!ot surveyed
Test #14 LOGGER available
Last survey at 14:44:51 on 06/05/73
5. DISPlJ\Yl rlG LO!jGT[r.I~ I:ISTORY OF A HOST
ULOIJGterl.1.I:IStory of tHC (you type "l on g<SP>hist<SP>nic<CR>i,)
. (You I,lay use any other acceptable host n arne or n umb e r )
PAGE 7
23232 Surveys taken beginning 19:43:24 on 04/27/72
Undetermined 0 ti~es ( 0%)
Host dead 5715 ti~es ( 25~)
NCP not responding 0 times ( O~)
LOGGER rejecting 154G ti r.es ( 7%)
LOGGER n o t responding 5 tines ( O?~)
LOGGER available 159G4 ti nes ( G9~)
Average response t I r.ie = 1. 70 seconds
Last survey at 10:03:32 on 06/05/73
G. DISPLAYJiJG HiE ACCEPTAGlE HOST I:Al iES
QH05T$ (to display the acceptable host names)
---J-- ---.------ I lOSTS ---------
DEC OCT STAiJDAHD IJAI·iE t·J I CK- :U\ll E5
U01.001 . uc 1 a nrnc
r sex-ucla
\)02 002 sri-arc nic
003 003 uc s b-t.io d 7 5
004 OU4 utah-10
005 005 bbn-ncc
DOG DOG m i t -1 .1u 1 tic s multics
007 007 rand-rcc
003 010 suc-adept
009 011 harv-10 harvard
010 012 11-G7
011 013 su-ai sail
012 014 ill-ants '
013 015 case-10
014 GIG cmu-cc cmu
015 017 i 11 i ac i4
DIG 020 ames-G7
017 021 r.l i t re
01S 022 radc-G45
019 023 ' n bs - c c s t
020 .0 24 etac
021 025 tink-418
022 02G l.lccl-418
023 027 usc-44
024 U30 gwc
025 031 noaa-7GOO
02G 032 saac
027 033 ai.ie c
023 U34 arpa
029 035 aberdeen
030 036 bbn-tip
031 037 cca
032 040 pa r c -rnax c
PAGE 8
033 041 f nwc
034 042 1b 1
035 043 ucsd-cc
065 101 ucla-ccn ccn-ucla
006 102 sri-ai
OG9 105 bbn-tenexa bbn tenex
070 lOG 1,1 it -dr'lcg dmcg its
071 107 rand-csg
073 III harv-1
074 112 ll-tx2
073 116 cmu-i l Oa l t
079 117 ame s
08G 12G usc-isi is i
09G 140 parc-vts
133 205 bbn-tenexb
131~ 20G r,lit-ai ai
137 211 harv-11
133 212 ll-tsp
1~7 305' bbn-1d
198 30G 1,1 i t -1.11
THE FOLLm:1 ilG HOSTS ARE TERf·lIl ·! AL Il~PS:
141~ 220 ar,le s - tip
145 221 r,li t r e - tip
14G 222 rauc-,tip
147 223 nbs-tip
143 .224 etac-tip
151 227 usc-tip
152 230 g\JC- tip
153 231 Jocb-tip
154 232 saac-tip
155 233 belvoir
15 ,0 234 a r o a-r t l o
158 236 bbn-testip
159 237 cca-tip
1G3 243 , a l o ha - t i p
IV. FUTURE SERVICES
\/e w l l l r.ia l n t a l n a CURREllT d a t ab a s e at the DATACm:UTER CCCA)
by transfering the survey l n f o rrna t l on to the Dt-TACOl tPUTEF as soon
as possible, l s e , , every 20 rn l nu t e s whe n e v e r the' f),\TACO/,iPUTEr; is
in operation. ~o it should be possible for other n e tv.o r k users
to wr l t e their O\Jn query p r og r am s in DAT;\LAt:GUAGE. \'!e \'Jill
provide one such Querry r ac l t l t v in our l:lDDLE s ub s y s t crn , He do
have an existing database of survey data at the O,\Ti,C01 :rUTr.r., anJ
also p r o g r ams exist in L LDDLE to access this database. As soon
as the use anti f o rr.ia t of the d a t a at the DATI\COr : i-'LJTEr. database is
PAGE 9
, , . .
- s t ab l 1 i zed we w l 11
publ ish a r.ierno on hOVI to use the database and
facilities to access it.
If it is considered desirable by nc two r k' users and ARPA, vie
w l l l provide specialized services to display survey data o n .
specific sockets, and/or via commands in fJET\JRK and f TDDLE. An
exa~ple of such special i zed services is to display a week's or a
-rnon t h f s Ho s t av a l l ab I l i t v profile.
v. REQUEST FOR COr-:1·1EIJTS AIm Sl.'GGESTI OI:S
. As r.icn t l on e d in m ·/G/RFC #523 (;llC #1 7 0 4 8 , IISUrUEY is in
Ope rat ion ;\ ;;a i n II) i f you des ire a c han 8 e i nth e l i s t of sit e s \'1 e
survey, please call rile at G17-253-1I~23 or send rna l l to akb Qr,lit
drnc g or at 141C (iJent= a k b ) , Please let mee know if survey is
l n corrv e n l e n c l n g anyone, \Je will try our best to resolve any
p rob 1 eras •
Your cornme n t s and suggestions are invited for I rnp r ov crue n t s
in the vav we collect, store, retrieve .. anJ display data. For
e xe i.ro l e , s hou l d we me a s u r e the response t l ri e for actually getting
s or.ic data on the Tt:LrlET connection, should we start s u r v e y l n g
FT P soc ke t 3 a S VI ell ( per hap 5 1c s s 0 f ten ), s h 0 u 1d \ l e t r y t 0
actually login (only occassionally at random) anJ collect
5 tat i s tic sun s y s t em loa din g , t i (;;eta ken to log i net c ., and \'! hat
services we should provide in addition to t ho s e vrne n t l on e d above.
AI ELI5
On this page
PROBLEMNetwork operators needed to know which ARPANET hosts were available and responsive at any given time, and users wanted to understand why a connection failed.
MECHANISMThe SURVEY project continuously tests connectivity to all known hosts on the ARPANET by attempting TELNET connections to the LOGGER socket (socket 1) every 20 minutes, recording the response time and status code for each host.
NETWORK VISIBILITY ENABLEDSURVEY provides a regularly updated network status service accessible via socket 15, storing results in a database at the DCC for analysis. Other hosts can query the data or display current host availability to users.
§1Introduction
RFC 530 reports on the SURVEY project, an automated program that regularly monitors ARPANET host availability. The SURVEY program is run by MIT and surveys every known "SERVER" host on the ARPANET at regular intervals (initially 15 minutes, later 20 minutes). It attempts to establish a full-duplex TELNET connection to the LOGGER socket (socket 1) on each host, measuring the time to establish the connection and recording the result.
MIT ran a program that tried to log into every computer on the network every 20 minutes. It wrote down which ones answered and how fast they answered.
§2The Survey Methodology
The SURVEY program tests each host by attempting an ICP (Initial Connection Protocol) to the LOGGER socket with a 5-second timeout per connection attempt. Response time is measured as the time to establish the full-duplex TELNET connection, serving as a lower bound on actual user response time.
The program records date, time, status code and response time. Status codes indicate: 5 = logger available, 4 = timeout, 3 = ICP aborted by foreign host, 2 = no reset response, 1 = host dead, 0 = not surveyed. Results are stored in ASCII format and transmitted to a database at the CCA (Communications Computer Access) DATACOMPUTER for centralized storage and analysis.
The test is simple: try to log in, time how fast it answers, record if it succeeded or why it failed. The results go to a big computer where they are saved so people can look them up.
§3Results and Services
The SURVEY results are published on socket 15 (decimal) on host 70 in a standard format showing month, day, year, hours, and minutes, followed by host number, status, and response time for each host surveyed. Individual hosts can periodically retrieve and store this information, display it to their users, or write their own query programs to access the SURVEY database at the DATACOMPUTER.
The RFC also mentions an online NETWORK program accessed via login to host 70, socket 1, which provides commands to display survey results, compute statistics across multiple surveys, and access long-term history. Future services planned include specialized displays for weekly or monthly host availability profiles.
Anyone could call socket 15 and get a list of which computers were up and how fast they answered. People wrote programs to store this data and show others which computers to use.