top - download
⟦4403c3d72⟧ Wang Wps File
Length: 22304 (0x5720)
Types: Wang Wps File
Notes: FIX/1154/DSP/0107
Names: »3173A «
Derivation
└─⟦7680f6628⟧ Bits:30006133 8" Wang WCS floppy, CR 0288A
└─ ⟦this⟧ »3173A «
WangText
…18……00……00……00……00…1…02……00……00…1
…08……08……08……0a……08……0b……08……0f……08……00……08…
…08… …08……06……08……07……07……0b……07……00……07…
…07… …07……05……07……86…1 …02… …02… …02…
…02…FIX/1154/PSP/0107
…02… JL/821217…02……02…
NODAL SWITCH SUBSYSTEM (NSS)
…02… FK 7809
3.4.4.3 T̲h̲e̲ ̲O̲u̲t̲b̲o̲u̲n̲d̲ ̲M̲e̲s̲s̲a̲g̲e̲ ̲B̲u̲f̲f̲e̲r̲
For each trunk and for each level of precedence the buffer contains a sub-buffer for 2 4-field
message-pointers (fig. 3.4.4.3-1):
1. MTCB index (1 word)
2. Serial No. (node-to-node) (1 word)
3. No. of retransmissions (1 - )
4. Routing Mask (2 words)
The elements are put into the buffer by the Transport Station before transmission of the
message (narrative + control excl. ACK/NACK); they are also removed by the Transport Station
after ACK or 4 NACK or 2 TIMER. The buffer is located in the memory; the messages are
stored on the disc.
The number of sub-buffers reserved each second is for an average Node (cfr. ch. 3.6.2):
(OUT…0f…n…0e…+OUT…0f…c…0e…)/8 = 2 1.97 /8 = 0.50 msgs/s.
all nodes
In average a sub-buffer is reserved for a period of time necessary for transmission and processing
the message and the reflected ACK:
Message waiting time in Trunk-Queue 20 s
Message transmission time (2.60 4.9) 13 -
ACK waiting time in Trunk Queue 2 -
A̲C̲K̲ ̲t̲r̲a̲n̲s̲m̲i̲s̲s̲i̲o̲n̲ ̲t̲i̲m̲e̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲2̲ ̲-̲ ̲ ̲ ̲
Total 37 s.
Using a safety factor of 4, the necessary buffer size is
0.50 37 4 = 7̲4̲ ̲s̲u̲b̲-̲b̲u̲f̲f̲e̲r̲s̲ per node
for message pointers, or app. 2 message pointers per trunk per level of precedence. Within
a precedence level the pointers are stored and removed in a cyclic way. The current state
of each sub-buffer is specified by FIRST-OCC and FIRST-FREE.…86…1 …02… …02… …02… …02…
…02…
FIG. 3.4.4.3-1:
O̲U̲T̲B̲O̲U̲N̲D̲ ̲M̲E̲S̲S̲A̲G̲E̲ ̲B̲U̲F̲F̲E̲R̲
OMB(0..T,S..R,0..(MW-l),1. .OMBFL)…86…W …02… …02… …02… …02… …02…
3.4.5 T̲a̲b̲l̲e̲s̲
3.4.5.1 T̲h̲e̲ ̲I̲/̲O̲-̲T̲r̲a̲n̲s̲f̲e̲r̲ ̲T̲a̲b̲l̲e̲
The I/O-Transfer Table contains one entry for each input- or output- operation which may
occur simultaneously on disc, trunks and LTUX…08…:
o Disc input : packets, files
o Disc output : - , -
o Trunk input : packets
o Trunk output : -
o LTUX Trunk : supervision
o LTUX Data User : -
o LTUX Crypto : -
see fig. 3.4.5.1-1.
An entry contains the TRANSFER-ID, also called the OP-REF, or #FFFF if the entry is free.
Further an entry contains the name of a semaphore and the type of an operation. The operation
is signalled to the semaphore, when the I/O-transfer is finished, i.e. when a system answer
from the I/O-SYSTEM is received by the Event Module.…86…W …02… …02… …02… …02…
…02…
FIG. 3.4.5.1-1:
THE I/O-TRANSFER-TABLE
3.4.5.2 T̲h̲e̲ ̲M̲e̲s̲s̲a̲g̲e̲/̲A̲n̲s̲w̲e̲r̲-̲T̲a̲b̲l̲e̲
The Message/Answer-Table contains one entry for each message which may be sent simultaneously
to the CHECKPOINT PROCESS awaiting an answer, see fig. 3.4.5.2-1.
An entry contains the MESSAGE/ANSWER-ID, also called the BUFFER-REF, or #FFFF if the entry
is free. Further an entry contains the name of a semaphore and the type of an opertion.
The operation is signalled to the semaphore, when the answer from the CHECKPOINT PROCESS
is received by the Event Module.
FIG. 3.4.5.2-1:
THE MESSAGE/ANSWER-TABLE
3.4.6 S̲e̲m̲a̲p̲h̲o̲r̲e̲s̲ ̲a̲n̲d̲ ̲O̲p̲e̲r̲a̲t̲i̲o̲n̲s̲
A total view of coroutines, queues, semaphores and operations of the NSS is found in fig.
3.4.6-1. The arrows indicate the flow of operations and queue-elements.
FIG. 3.4.6-1:
COROUTINES, QUEUES,
SEMAPHORES AND OPERATIONS
3.4.6.1 G̲e̲n̲e̲r̲a̲l̲ ̲L̲a̲y̲o̲u̲t̲ ̲o̲f̲ ̲O̲p̲e̲r̲a̲t̲i̲o̲n̲s̲
The small piece of information signalled from one coroutine to another via a chained semapore
is called an o̲p̲e̲r̲a̲t̲i̲o̲n̲. The specific contents are described in the chapters 3.3.4 through
3.3.9; however, the general layout is specified below:
B̲Y̲T̲E̲ ̲#̲ ̲ N̲O̲.̲O̲F̲ ̲B̲Y̲T̲E̲S̲ C̲O̲N̲T̲E̲N̲T̲S̲
1-0 2 succ
3-2 2 prec
5-4 2 s̲e̲n̲d̲e̲r̲: module
7-6 1+1 d̲e̲v̲i̲c̲e̲ # or q̲u̲e̲u̲e̲ #,
t̲y̲p̲e̲ of operation
9-8 2 i̲n̲f̲. (1,0)
11-10 2 i̲n̲f̲. (3,2)
T̲h̲e̲ ̲s̲e̲n̲d̲e̲r̲ may be:
Packet Handler
Transport Station
Monitoring Module
Control Module
Event Module
Starting Module
T̲h̲e̲ ̲t̲y̲p̲e̲ may be:
Input
Output
Status
Commnd Response
Timer
etc.
T̲h̲e̲ ̲d̲e̲v̲i̲c̲e̲ # may be the LTUX #.
The balance contains additional i̲n̲f̲o̲r̲m̲a̲t̲i̲o̲n̲.
3.4.7 C̲o̲n̲s̲t̲a̲n̲t̲s̲ ̲a̲n̲d̲ ̲V̲a̲r̲i̲a̲b̲l̲e̲s̲
Important message transmission constants are:
NPL: No. of Precedence Levels.
MTIMER: The value in seconds of the message-timer.
MRETRANS: The maximum number of retransmissions.
MAXMSG: The maximum length of a message.
MW: The window
Important packet transmissions constants are:
PACKSIZE: The maximum number of databytes.
PW: The window.
The medename contains the name of the MEDE or MEDE…08…s if more than one., f.ex. : KZ: .
In fact the name is duplicated: INAME and ONAME. During test they may be set to different
values f.ex. INAME:= :K: and ONAME:= :Q: simulating two nodes in a single CR80.
Local Networks Constants may be:
T: The number of trunks.
DU: The number of data users.
FIKS-STATE is a variable, which currently tells the state of the FIKS network: normal, alternate
1, or alternate 2.
3.5 S̲t̲o̲r̲a̲g̲e̲ ̲A̲l̲l̲o̲c̲a̲t̲i̲o̲n̲
This chapter describes the primary storage requirements (program and data in memory) and
the backing storage requirements (files on disc).
3.5.1 M̲e̲m̲o̲r̲y̲ ̲S̲p̲a̲c̲e̲ ̲R̲e̲q̲u̲i̲r̲e̲m̲e̲n̲t̲s̲
3.5.1.1 P̲r̲o̲g̲r̲a̲m̲ ̲a̲n̲d̲ ̲D̲a̲t̲a̲
The space occupied by the data is strongly dependant on the number of trunks T radiating
from the node for saving space on the small nodes.
The space occupied by the program is independant of the node, except for the SCC…08…s where small
pieces of code are omitted (cfr. appendix 7.2).
The space is calculated from fig. 3.5.1-1 (node K), and fig. 3.5.1-2 (nodes P and Q).
The result is shown in fig. 3.5.1-3.
FIG. 3.5.1-1:
NSS MEMORY LAYOUT
NODE Q (T=1)
FIG. 3.5.1-2:
NSS MEMROY LAYOUT
NODE K (T=7)
3.5.1.2 M̲e̲s̲s̲a̲g̲e̲ ̲B̲u̲f̲f̲e̲r̲s̲
The buffers are used for Kernel Messages/Answers exchanged with
- The TIMER process
- the CHECKPOINT process
- the I/O-System
The No. of Kernel Messages for the T̲I̲M̲E̲R̲ ̲p̲r̲o̲c̲e̲s̲s̲ is:
- 1 Kernel-msg for call "on the hour" (1 hour)
- 1 Kernel msg for LTUX-polling ( 30 s)
- 1 Kernel msg for each transmitted
un-acknowledged FIKS message ( 50 s).
For a Busy Hour Traffic in node K: OUT = 0.676
narrative and control msgs/s (fig. 3.6.2.1-3)
and using a safety factory of 3 (busy minute)
the average No. of TIMER msgs is:
2 + 50 0.676 3 = 100 msgs.
The No. of Kernel msgs for the C̲H̲E̲C̲K̲P̲O̲I̲N̲T̲ ̲p̲r̲o̲c̲e̲s̲s̲ is:
(see ch. 3.3.10.3):
- 1.7 Kernel msg/s
each occupied in the average for 50 ms. Using a safety factor of 9 (busy second), the average
No. of CHECKPOINT msgs is:
0.050 * 1.7 * 9 = 0.9 1 msg.
The No. of kernel messages for the I̲/̲O̲-̲S̲y̲s̲t̲e̲m̲ has an upper limit equal to the number of entries
in the I/O-Transfer Table:
62 msgs.
The total No. of message buffers is:
100 + 1 + 62 = 163 msg buffers
each occupying 5 words. The buffers are allocated from a common process pool, and therefore
they are not part of the NSS base.
3.5.1.3 F̲i̲l̲e̲ ̲D̲e̲s̲c̲r̲i̲p̲t̲i̲o̲n̲s̲
The File Descriptions are used for:
- File Directories
- Files permanently open
- Files with FIKS messages being processed
by the NSS
- Loglines to the LTUX…08…
The File Directories used are:
- FIXHEAD, MOVHEAD and PDB.D
or a total of 3.
The files permanently open are:
- HDB (1)
- IMF (2)
- NDF (1)
- CPF (1)
- LCF (2)
or a total of 7.
The LCF…08…s are used during configuration only.
The No. of PDB messages being processed by the NSS during busy hour in node K (OUT = 0.676
msgs/s, fig.3.6.2.1-3) is in average using a safety factory of 9 (busy second):
0.676 4.6 2.6 9 = 66.3 67
using a transmission time of 4.6 s/packet (cfr. ch.3.3.11.1) and an average message length
of 2.6 packets/msg (appendix 7.1).
Finally the No. of Loglines to the LTUX…08… are calculated for Node K:
LTUX SUB.DEV.#2 #3
TRANS 1 1
TRUNK…08…s 6 0
NPDN 1 0
DATA USER…08…s 7 0
CRYPTO, RED 1 1
- , BLACK 0 0
̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲
TOTAL 16 + 2 = 18
̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲
So the total No. is:
3 + 7 + 67 + 18 = 95 file descriptions
each 18 words long.
3.5.1.4 I̲/̲O̲-̲C̲o̲n̲t̲r̲o̲l̲ ̲B̲l̲o̲c̲k̲s̲
The max. No. of IOCB…08…s is determined in this way (cfr. the I/O-Transfer Table):
Msg. from disc to CM 1
Packets from disc to PH 2
Msg. from MM to disc 1
Packets from TSI to disc 7
OMB from TSO to disc 1
Packets from trunks to PI 14
Packets from PO to trunks 1
Polling of LTUX-TRK…08…s 7
- - -DU…08…s 25
- - -Crypto…08…s 1
̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲
Max. No. of IOCB's 60
each 35 words long.
3.5.1.5 T̲r̲a̲n̲s̲f̲e̲r̲ ̲L̲i̲s̲t̲ ̲E̲l̲e̲m̲e̲n̲t̲s̲
The No. of TLE…08…s equals the No. of IOCB…08…s i.e. 60 TLE…08…s, each occupying 5 words.
3.5.2 D̲i̲s̲c̲ ̲S̲p̲a̲c̲e̲ ̲R̲e̲q̲u̲i̲r̲e̲m̲e̲n̲t̲s̲
The HDB, PDB and the LCF…08…s are used by other subsystems also; their size is described in ref.
4.
The size of the IMF is calculated in ch. 3.4.2.3.1. For all nodes it is:
1 032 192 bytes.
Refering to ch. 3.4.2.4 the size of the NDF is:
1 024 bytes.
The length of the Jack File is (ch. 3.4.2.5):
192 bytes.
The length of the Checkpoint File (or the Outbound Message File) depends on the number of
trunks, cfr. ch. 3.4.2.7:
2 + (T+1) 140 bytes.
3.6 P̲e̲r̲f̲o̲r̲m̲a̲n̲c̲e̲ ̲C̲h̲a̲r̲a̲c̲t̲e̲r̲i̲s̲t̲i̲c̲s̲
3.6.1 E̲q̲u̲a̲t̲i̲o̲n̲ ̲o̲f̲ ̲C̲o̲n̲t̲i̲n̲u̲a̲t̲i̲o̲n̲
The increase of the number of messages p̲e̲r̲ ̲u̲n̲i̲t̲ ̲o̲f̲ ̲t̲i̲m̲e̲ in a node is:
dm
-- = IN + ORIG + GEN - OUT - DEST - DEL
dt
where
IN: No. of messages inbound from the trunks
ORIG: - - - originating from the
MEDE/SIP
GEN: - - - generated incl. copying in
the node
OUT: - - - outbound on the trunks
DEST: - - - destinating the MEDE/SIP
DEL: . . . deleted in the node
During s̲t̲e̲a̲d̲y̲ ̲s̲t̲a̲t̲e̲ conditions:
dm
-- = 0
dt
so t̲h̲e̲ ̲e̲q̲u̲a̲t̲i̲o̲n̲ ̲o̲f̲ ̲c̲o̲n̲t̲i̲n̲u̲a̲t̲i̲o̲n̲ is valid:
IN + ORIG + GEN = OUT + DEST + DEL (1)
cfr. fig. 3.6.1-1.
No messages are generated nor deleted on the trunks, so:
IN = OUT (2)
all nodes all nodes
From (1) and (2) one finds:
ORIG + GEN = DEST + DEL (3)
all nodes all nodes all nodes all nodes…86…W …02… …02… …02… …02… …02…
…02…
…01…FIG. 3.6.1-1:
FOR A NODE DURING STEADY STATE CONDITIONS
THE "EQUATION OF CONTINUATION" IS VALID:
IN + ORIG + GEN = OUT + DEST + DEL
3.6.2 N̲o̲d̲e̲ ̲T̲r̲a̲f̲f̲i̲c̲
3.6.2.1 N̲a̲r̲r̲a̲t̲i̲v̲e̲ ̲M̲e̲s̲s̲a̲g̲e̲s̲
In fig. 3.6.2.1-1 and 3.6.2.1-2 is shown the effect of putting a (single address) narrative
message for each remaining MEDE into the network, or putting one seven-address message into
the network. The 8 nodes are interconnected as shown, but we do not make any assumptions
about the exact No. of trunks between two neighbour nodes.
The numbers of the left hand column and the right hand column of fig. 3.6.2.1-2 have been
added; the results are shown in fig. 3.6.2.1-3a and b, the last one after being divided by
7, so they are both valid for one input message per node.
The network multiplicity is 2.3. This may for example be achieved by mixing 21.7% seven-address
messages and 78.3% one-address messages, because:
0.217 7 + 0.783 1 = 2.30
Adding table a and b in this way and dividing by 8 for normalizing (1.00 input message for
the entire network), one gets fig. 3.6.2.1-3c. So we assume an equal number of originating
messages entering each node, and an equal number of incoming messages leaving each node.
In average 86% of the inbound narrative traffic will have arrived its destination, because:
DEST…0f…n…0e… 2.300
------- = ----- = 0.86
IN…0f…n…0e… 2.664
During busy hour 2692 narrative messages enter the total network, which corresponds to
2692/3600 = 0.7478 narr.msgs/s.
By multiplying fig. 3.6.2.1-3c by this factor one gets fig. 3.6.2.1-3d: Narrative messages
during busy hour.…86…W …02… …02… …02… …02… …02… …02…
FIG. 3.6.2.1-1:
ROUTING OF MESSAGES
FIG. 3.6.2.1-2:
7- AND 1- ADDRESS MESSAGES
THROUGH THE NETWORK
FIG. 3.6.2.1-3:
3.6.2.2 C̲o̲n̲t̲r̲o̲l̲ ̲M̲e̲s̲s̲a̲g̲e̲s̲
Control messages are now added to the model. First we consider control messages other than
AKC/NACK…09…s, and finally ACK…08…s and NACK…08…s are also added.
Experiences until now seems to indicate an idle load of control messages other than ACK/NACK…08…s
equal to the busy hour load from narrative messages when counting number of messages. This
load is shown in fig. 3.6.2.1-3e.
3.6.2.3 A̲K̲C̲/̲N̲A̲C̲K̲…88…s̲
For each narrative or control message sent to a neighbour node, one ACK (or NACK) will be
received and vise versa, i.e.:
IN…0f…a…0e… = OUT…0f…n…0e… + OUT…0f…c…0e… (4)
OUT…0f…a…0e… = IN…0f…n…0e… + IN…0f…c…0e… (5)
ACK/NACK…08…s are not sent to or from the MEDE/SIP:
ORIG…0f…a…0e… = DEST…0f…a…0e… = 0 (6)
From (1), (2) and (6) is found:
GEN…0f…a…0e… = DEL…0f…a…0e… (7)
all nodes all nodes
i.e. the ACK/NACK…08…s generated inside the network are also deleted inside the network.
Disregarding alternate routing all outbound ACK/NACK…08…s are generated by the node:
OUT…0f…a…0e… = GEN…0f…a…0e… (8)
and all inbound ACK/NACK…08…s are deleted by the node:
IN…0f…a…0e… = DEL…0f…a…0e… (9)
see fig. 3.6.2.1-3f
3.6.3 T̲i̲m̲i̲n̲g̲ ̲o̲f̲ ̲S̲u̲b̲s̲y̲s̲t̲e̲m̲ ̲E̲x̲e̲c̲u̲t̲i̲o̲n̲
The calculation of the C̲P̲U̲-̲l̲o̲a̲d̲ ̲f̲r̲o̲m̲ ̲t̲h̲e̲ ̲N̲S̲S̲ ̲is performed for node K, which is the node of
heaviest load during busy hour.
The traffic consists of:
a. Narrative and Control messages:
- inbound messages for the MEDE
- inbound messages for the SCC
- relay messages
- outbound messages from the MEDE
- outbound messages from the SCC
- messages generated by copying
b. ACK…08…s and NACK…08…s:
- inbound ACK/NACK messages
- outbound ACK/NACK messages
The a̲v̲e̲r̲a̲g̲e̲ ̲C̲P̲U̲-̲t̲i̲m̲e̲s̲ used are:
- Instruction time: 2.2 us
- QACCESS: 0.2 ms
- MTCB access: 0.2 -
- Trunk I/O: 3.2 -
- Disc I/O: 3.2 -
The calculation is made in fig. 3.6.3-1. The CPU-load from the NSS is found to be 3̲%̲, mainly
from trunk I/O.
FIG. 3.6.3-1:
CALCULATION OF THE CPU-LOAD FROM THE NSS
THE TOTAL LOAD IS 3%
3.7 L̲i̲m̲i̲t̲a̲t̲i̲o̲n̲s̲
This page is intentionally left blank.…86…W …02… …02… …02… …02… …02… …02…
3.8 E̲r̲r̲o̲r̲ ̲C̲o̲d̲e̲s̲/̲E̲r̲r̲o̲r̲ ̲L̲o̲c̲a̲t̲i̲o̲n̲s̲
The belowmentioned errors are those which may occur in the error free NSS Software. The error
locations are spread all over the NSS, and the listing may be consulted. Via the AMOS KERNEL
procedure ERROR the ESP is called, after which either a return is made to the NSS for a
local fix-up, or the entire N/M is switched over to the standby branch.
3.8.1 K̲e̲r̲n̲e̲l̲
#0101 Trap Serious error in another substem, or in a file. Consult the
NSS listing. N/M Switchover.
#0102 Parity Error in N/M Switchover.
memory or on bus.
3.8.2 L̲T̲U̲X̲ ̲I̲/̲O̲
#0607 Time out during I/O. Local Fix-up.
#0B01 Time out when The location field
polling a LTUX contains the bus and the device #:
black 15+dev 8+0
Local Fix-up.
#0B02 No answer from The location field
a LTUX 30s after contains the bus and
polling. the device #:
black 15 + dev 8+0
Local fix-up.
3.8.3 Q̲A̲C̲C̲E̲S̲S̲
#0801 No free queue element Local fix-up.
#0805 Queue overflow Local fix-up.
3.8.4 M̲T̲C̲B̲-̲M̲o̲n̲i̲t̲o̲r̲
#0907 No MTCB available Local fix-up
#090B No IMF area avail. Local fix-up
3.9. L̲i̲s̲t̲i̲n̲g̲ ̲R̲e̲f̲e̲r̲e̲n̲c̲e̲s̲
Ref. to SOURCE LIBRARY.
4 Q̲U̲A̲L̲I̲T̲Y̲ ̲A̲S̲S̲U̲R̲A̲N̲C̲E̲
4.1 Q̲u̲a̲l̲i̲f̲i̲c̲a̲t̲i̲o̲n̲ ̲T̲e̲s̲t̲s̲
The follwing tests were performed:
- Procedure Tests
- Module Tests
- Subsystem Test
- N/M System Test
- Collocated N/M System Test
- SCC System Test
- Red TDX Test
- Black TDX Test
- TRANS Test
- Micro System Test
- Trunk Test
- NPDN Test
- Data User Test
- Micro Network Test.
They are all described in ref. (12).
4.2 O̲t̲h̲e̲r̲ ̲Q̲u̲a̲l̲i̲t̲y̲ ̲A̲s̲s̲u̲r̲a̲n̲c̲e̲ ̲P̲r̲o̲v̲i̲s̲i̲o̲n̲s̲
The belowmentioned tests were also run before the Mini Acceptance Test:
- Load Test
- Long Term Test
- Stress Test
- Roughness Test
described elsewhere.
5 P̲R̲E̲P̲A̲R̲A̲T̲I̲O̲N̲S̲ ̲F̲O̲R̲ ̲D̲E̲L̲I̲V̲E̲R̲Y̲
Preparation of a binary NSS for a NODE/MEDE starts with an editing of the source text. As
a minimum the name of the MEDE (…08…s) must be set in the HEAD, see fig. 5-1.
Then the source texts incl. the HEAD and the TAIL are merged together with the AMOS-G1 and
-G2 text files, the control-message declarations, the External Declarations, and the Coroutine
Monitor.
The Merge file is assembled by the LARGE-ASSEMBLER, and the listing is printed.
Finally the binary NSS so produced may be installed.
FIG. 5-1:
PROGRAM MAINTENANCE
6 N̲O̲T̲E̲S̲
This page is intentially left blank for Your notes.
7 A̲P̲P̲E̲N̲D̲I̲C̲E̲S̲
7.1 T̲h̲e̲ ̲L̲e̲n̲g̲t̲h̲ ̲o̲f̲ ̲M̲e̲s̲s̲a̲g̲e̲s̲
Referring to figure 7.1-1 the mean value of the length of narrative messages is:
9072
̲ 1̲
1 = 1 F (1) dl = l s e 1…0f…1…0e…dl
0 1= 0
or
-9072/1…0f…1…0e…
1000 = s 1…0e…2…0f… x e…0e…x…0f… dx s 1…0e…2…0f… (1)
…0e…1…0f… …0e…1…0f…
x=0
One also finds:
9072
̲ 1̲
1 = F (1) dl = s e 1…0f…1…0e… dl
0 l=0
or 9072/1…0f…1…0e…
1 = sl…0f…1…0e… e…0e…-x…0f… dx s 1…0f…1…0e… (2)
x = 0
so from (1) and (2)
1…0f…1…0e… 1000 chars
̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲
-3
s̲ ̲ ̲ ̲ ̲ ̲ ̲1̲ ̲0̲0̲0̲ ̲ ̲ ̲1̲0̲ ̲ ̲ ̲ ̲
The distribution of packets is found in fig. 7.1-2.
The average message length is found to be:
̲
p̲ ̲=̲ ̲2̲.̲6̲0̲ ̲p̲a̲c̲k̲e̲t̲s̲/̲m̲e̲s̲s̲a̲g̲e̲
FIG. 7.1-1:
THE DISTRIBUTION
OF MESSAGE LENGTHS
FIG. 7.1-2:
THE DISTRIBUTION
OF PACKETS
7.2 T̲h̲e̲ ̲N̲S̲S̲ ̲i̲n̲ ̲t̲h̲e̲ ̲S̲y̲s̲t̲e̲m̲ ̲C̲o̲n̲t̲r̲o̲l̲ ̲C̲e̲n̲t̲e̲r̲s̲
Each System Control Center (i.e. SCC P and Q) has the NSS installed communicating with the
NSS in the collocated node.
The small differences between the NSS in the SCC…08…s and the NSS in the other nodes are explained
below.
The traffic transmitted between a SCC and the collocated node is n̲o̲n̲-̲e̲n̲c̲y̲p̲t̲e̲d̲; therefore,
there are not LTUX-CRYPTO…08…s and the Packet Handler doesn…08…t manage a log line for encrypted
data.
There are n̲o̲ ̲D̲a̲t̲a̲ ̲U̲s̲e̲r̲s̲ and consequently no Jack File.
There is no LTUX polling.
The TRANS-connection between the SCC and the collocated node can n̲o̲t̲ be replaced by a N̲P̲D̲N̲
̲c̲o̲n̲n̲e̲c̲t̲i̲o̲n̲.
The queue corresponding to the MDS-Q is called the C̲I̲P̲-̲Q̲.
N̲o̲d̲e̲ ̲R̲e̲s̲t̲a̲r̲t̲ ̲w̲i̲l̲l̲ ̲n̲o̲t̲ ̲b̲e̲ ̲p̲e̲r̲f̲o̲r̲m̲e̲d̲. Therefore there are no checkpoints and the Outbound Message
Buffer is not dumped.
C̲o̲n̲t̲r̲o̲l̲ ̲M̲e̲s̲s̲a̲g̲e̲s̲ similar to those generated at other nodes for the MEDE Supervisors are n̲o̲t̲
generated on the SCC…08…s.
