GTE Omni Si Database Technical Practices Issue 1 Manual

							TL-130500-1001
System Capacities12.3 Table 12.3 provides slot requirements and PLA (Packet
Line Address) requirements for data card and remote devices.
Table 12.3PCB/Device Quantity per System
CARD OR DEVICEQUANTITYNO. OF
PCMUS SLOTSADMPA&CFB-17229-A FB 
17230-BOA
UCB/DCP
FB-17231 -A1 per system
1 per system1 each
(2 total)
1Packet Router
FB-17228-BOA1 per system
1
PBE/T
FB-17227-A
VPLC (type VPLO)
FB-17226-AThis # depends on
the packet bus
configuration.
1 per 8 ports
VPLC (type 
VPLl)FB-17226-A
VPLC2 (type 
VP20)FB-17246-A
VPLC2 (type VP21)
FB-17246-AAPM
SPM
NCFB-17242-A
DFP/APM1 per 2 ports
11 per 8 voice +
data ports
1 per 2 voice 
+data ports
1 per
asynchronous port
1 per X.25 port
16 per system
1 per
asynchronous port
(also with voice
requirements)
i -.SVR 5210
8187s-1 93 
						

							TL-130500-1001Power Requirement12.4 Table 12.4 provides power requirement for data cards and
remote devices.
Table 12.4Physical Location and Power Requirements of Cards
S-l 94TYPEESTIMATED POWER REQUIREMENTS
+ 5VDC+ 12VDC-12VDC-48VDC1 
VPLCVPLC210.834 Amp
1  UCB (see Note)12.62 Amp0.038 Amp 
-IADMP (2 cards)
4.48 Amp0.13 Amp0.056 Amp 
-
PR2.2 Amp0.038 Amp -
PBE/T (see Note) -RPTR
1.5 Amp
1  
NCI2.0 Amp
APM
Not powered from the switch
SPM
Not powered from the switch
NOTE: When these cards are equipped with terminators, they
will require an additional 0.08 ampere at + 
SVDC.Power Supply Limits
12.5 The power supply limits of the system can be found in the
TL-130000-1001, paragraph 2.12.
System12.6 The following card placement is recommended for data
Recommendationcards.
e Place the ADMP A & C cards in the Get Started File. The
ADMP A& C cards must be in two adjacent slots.
e Place the UCB in the Get Started File.
0 Configure the UCB as a UCB/T (bus terminating) by placing it
in an end PCMUS.
@ Place the PR in the Get Started File.
8187SVR 5210 
						

							TL-130500-1001Local Packet Bus12.7 The general configuration rules for the data system can be
Configurationcharacterized as falling into one or more of the following
Rulescategories:
l 
PHI’ (Physical Limitation) = limited by physical space or
number of slots.
l POW (Power/Thermal) 
- limited by available power or thermal
considerations.
l SYS (System) = system design limitation
l MEM (Memory) = available memory limitation
l PRO (Processor Power) = software processing power
limitation
l CON (Configuration) = general configuration rule
Each of the following rules specifies within parentheses which of
the above categories is causing the rule to be in effect:
l The PEC can support one packet router enhanced (PHY,
POW).l A packet router can directly support two local packet buses
(SYS).l Each UCB and ADMP requires one PLA.A VPLO requires
eight 
PLAs, while VPLi requires two PLAs. Packet routers and
packet bus extender/terminators do not require 
PLAs (SYS).
NlCs require 1 PLA.
l Each UCB, NIC, VPLC, PR, and 
PBE/T card requires one
PCMUS, while the ADMP cards require two adjacent slots, only
the A side of which requires power (SYS, PHY).
l Each asynchronous port requires its own APM or DFP/APM
(SYS).l Each X.25 line which runs less than or equal to 19.2 Kbps can
use one SPM with the RS-232-C interface. The V.35
interface (SYS) supports X.25 lines which runs up to 64 Kbps.
l The system supports a maximum of 127 DFP (PRO, MEM).
l A single UCB card can support up to 240 data ports (MEM,
PRO).l A UCB card can support 240 data ports (for future release)
The UCB software is only required if data is present (SYS,
PRO, MEM).
SVR 5210
Em7s-1 95 
						

							TL-130500-1001
l A VPLC2 type VP20 card will support up to eight DFPs (with or
without the APM), 
APMs, or SPMs running at not greater than
19.2 2Kbps in any combination. Alternatively a VPLC2 type
VP21 card can support one SPM running at up to 64 Kbps and
one other device (SPM at any speed or APM) (PRO, PHY).
e A PEC has two files (shelves) (PHY,SYS).
* The total number of VPLC, ADMP, and VCIP cards per local
packet bus primary or extension is eight (POW, PRO).
l No data card can reside in the last PCMUS of group (CON).
l The maximum number of 
UCBs allowed per system is 1 (POW,
PRO).l Each active local packet bus must be terminated with either a
UCBIT card or a PBT card (SYS).
l A packet router requires a PBE card to gain access to its
second local packet bus (SYS).
General Rules12.8 The following general rules apply when configuring a data
system.
l The system contains 36 
PCMUSs (universal slots) suitable for
data use. Due to the size and organization of the port tables,
slot Al 1, 
Bll, Cl 1, Dll cannot be used by the data system.
Slot 
A0 is not used by data cards.
l One PR (Packet Router) can be connected to a maximum of
one LPB (Local Packet Bus).
l PR supports up to 64 PLA (Packet Line Addresses).
l To gain access to additional 
PLAs, the PRE must connect to a
second LPB (LPB 1). This connection is through the PBE
card.
When it is set, a switch on the UCB allows the card to function
as a terminator. If the switch is not set, the card does not
function as a terminator. In that case, the card can be placed
between a PR or PBE and PBT.
S-196
8187SVR 5210 
						

							SVR 5210l The PR (Packet Router) routes mini-packets on up to two
LBPs. The first LPB (LPB-0) on a PR is bounded by the PR
and a PBT (Packet Bus Terminator), or UCB/DCP with Bus
Terminator (UCB/BT). The second LPB (LPB-1) is bounded
by a PBE (Packet Bus Extender) and a PBT.
Twelve hundred mini-packets a second, per LPB, can be
routed through the PRE without overflowing.When configuring
the system, traffic calculations should be made accordingly.
The bits-per-second rate of all asynchronous and
synchronous devices should not exceed 600,000 bits per
second. For example, if there are 11 
SPMs operating at 64
Kbps each, then they should not be configured on the same
LPB.The following paragraph provides traffic information for
the data system.
Data Traffic12.9 Data terminal output information is linked to the CPU
Considerations(Central Processor Unit) by interface devices operating at
different data rates. Compatibility between the interface device
and PABX allows for uninterrupted transfer of data. Incompatible
interface devices require protocol conversion equipment. Data
rates with varying ranges are used to transmit data over a
specified communications line. The type of data transmitted can
fall into one or more of the following categories:
l Typed words
l Numbers
l Graphics
l Syntax
l Bit sequences
l Files
The primary considerations for transmission of any category
depends upon the user’s PABX configuration, type of link, and
network interfaces. Protocol requirements control the physical
and near real-time aspects of data transfer, and therefore must
be the primary software consideration.
Programmable terminals and host 
CPUs transfer data formats in
either a synchronous or asynchronous mode of transmission.
The format of the data, as it appears on the communications link,
may or may not arrive at its destination in the most economical
manner. Thus, the efficiency of data transmission depends upon
throughput rates, peak load periods, and the specific format.
8187s-1 97 
						

							- -TL-130500-1001
Data Throughput
12.9.1 For each APM, SPM, or 
DFP/APM connected to the
Considerationssystem, the PD-200 Data System supports either the
combination of one voice communication and up to 19.2 Kbps
asynchronous data communications, or one voice
communication and up to 64 Kbps synchronous data
communications. Throughput on each LWL (Local Wire Loop)
has been sized to accomplish this. Each LWL working in 
half-duplex mode at 256 kilobytes can transfer 1,330 mini-packets
per second in each direction. Table 12.5 shows the 
mini-packet equivalent for each throughput specification. The 64
kilobyte synchronous calculation includes the overhead for data
control, while the overhead for voice and data control is included
in the 19.2 kilobyte asynchronous calculation.
Table 12.5
Mini-Packet Equivalent for Each Throughput SpecificationFUNCTIONMINI-PACKETS/SECOND
S-1 988187SVR 5210 19.2 kilobytes asynchronous
I
300I
64.0 kilobytes PCM (Voice)
64.0 kilobytes synchronous
1,000
1,200Hardware on the VPLC strips off receive voice mini-packets and
merges them into the voice switch (time-division multiplex)
domain. Only non-transparent mini-packets reach the PTS
(Packet Transport System). These mini-packets include voice
control messages, data control messages, and X.25 packet
messages. For each asynchronous port (or synchronous port at
19.2 Kbps or less), the mini-packet traffic is limited to 300
mini-packets per second. The VPLC is sized to handle eight
ports, each running at 300 mini-packets per second. For high
speed data service, the VPLC is limited to running 2 ports at a
rate of 1,200 mini-packets per second (data rate of 64 Kbps).
Traffic Metering12.9.2 To ensure that congestion does not occur within the
VPLC, all MPP (Mini-Packet Protocol) end-point components
in the system can restrict the rate at which mini-packet traffic is
generated to a given destination. The restricting process is
called metering. When a component such as a 
UCB/DCP card
sends a message through a VPLC to an APM, the UCB card
sends the message mini-packets at a rate of one mini-packet
each 3.3 milliseconds (or less). Similarly, two high-speed SPM
managers communicating can send mini-packets to each other
at a rate of one mini-packet per millisecond. 
						

							SVR 5210TL-130500-1001
.If multiple components send messages to the same destination
component simultaneously, a temporary overflow condition can
occur. Mini-packet buffers in the VPLC and in the 
PRs are
used to hold the overflow. If the overflow condition persists, the
data system invokes an appropriate flow control mechanism.
Flow Control12.9.3 Flow control is inherent in the design of the data system
and exists at both the message and the mini-packet levels. It
provides speed matching for communicating subscriber devices
running at different data rates, and is also used to recover from
internal data congestion situations. Both levels of flow control are
invoked automatically as required.
Speed Matching12.9.4 All user data within the switch is sent by using the X.25
Flow Controlpacket level procedure. The packet window rotation mechanism
(number of unacknowledged packets which may be outstanding)
regulates traffic between two users.
For example, the standard window for an APM is two X.25
packets. If a 64 Kbps X.25 host sends packets to a 300-baud
terminal attached to an APM, flow control is used to provide
speed matching as follows:
l A packet is received at 64 Kbp’s from the X.25 host by the
SPM.
l The packet is sent as a series of mini-packets metered at a
rate of 300 mini-packets per second.
l The packet is reassembled at the APM. If it has been received
error free, the user data is sent to the terminal at a rate of 300
baud.l The X.25 packet level acknowledgment is sent to the SPM, and
then on to the X.25 host only after the entire packet has been
transmitted to the user by the APM.
l In this example, a second packet may be sent to the APM
while the first is being serviced. It is queued by the APM.
According to X.25 standards, the X.25 host can send no more
packets to the APM because the window is full (two packets
are unacknowledged).
In addition to the above rules, X.25 allows the user or the data
switch to send RNR (Receiver Not Ready) packets to enable flow
control without having to wait for a full window condition. The
data system does not generate 
RNRs but accepts them when
received from an X.25 subscriber.
8187S-l 99 
						

							TL-130500-1001Congestion12.9.5 Congestion can occur within a normally operating data
Flow Controlsystem for two reasons:
l Due to the “bursty” nature of data traffic, a fully configured
system may saturate the internal bus structure temporarily.
l Many end points within a data system may be communicating
with a single end point (i.e., a group of 
APMs with X.25 calls
established through the same SPM). It is statistically possible
that if enough end points send simultaneously, the path to the
destination end point will become congested.
The MPP protocol detects a congestion condition and recovers
from it. The mechanism works as follows:
l Congestion is detected when mini-packets are lost in
transmission. The MPP receivers do not acknowledge invalid
sequences received.
l When an acknowledgment is not received, MPP sends a
time-out and then schedules the message for retransmission.
e To prevent a second congestion from occurring, a pseudo-random timer is run at each sending MPP before resending the
message. This is called adaptive 
backoff.l Adaptive 
backoff is applied repetitively, with longer time-out
periods, until messages are sent successfully. Adaptive
backoff then gradually allows normal operation to resume.
This action is called damping.
l If adaptive 
backoff cannot handle the congestion, logical links
are marked down and calls are cleared until the traffic load
within the switch goes under the congested level. This
happens only if the data system is over-configured.
Configuration12.9.6 Table 12.6 shows the maximum throughput in full-
Limitsduplex mini-packets per second for each MPP end point in the
data system.
Table 12.6
Mini-Packets/Second for MPP End Pointss-200MPP END POINTMINI-PACKETS/SECONDI
IAdministrative/Maintenance Processor
1,000Universal Controller Board card
1,000Synchronous Packet Manager 
(> 9.6 kbps)
Synchronous Packet Manager 
(< 9.6 kbps)
1,200
300Asynchronous Packet Manager
300
8187SVR 5210 
						

							TL-130500-1001
Examine the figures in Table 12.6 versus those in Table 12.7,
which shows the mini-packets/second capability of the VPLC
and the PTS (Packet Transport System).
Table 12.7Mini-Packets/Second Capability of the VPLC/PTS
VPLC/PTSMINI-PACKETS/SECOND
VPLC (VPLO)2,400 (8 X 300)VPLC (VPLl)2,400 (2 X 1,200)
LPB (Local Packet Bus)
12,000PR (Packet Router)24,000 (2 X 12,000)I
Observe that each LPB (Local Packet Bus) is limited to 12,000
mini-packets/second. This limit must be taken into account
when planning installation of a fully configured system. This
prevents one LPB or one PR (Packet Router) from becoming
congested. It is normal for a data system to use an average
bandwidth far less than the maximum bandwidth available. The
throughput of the LPB and PR takes this into account. Thus, if
the subscriber devices burst data or run at a rate of one-half the
maximum (or less), then any LPB can be selected for the
connections. With maximum throughput devices, the limit is 40
asynchronous devices or 9 synchronous devices per LPB.
Data Capacity12.9.7 The data switch can handle a maximum of 120 data ports
Considerationsoperating at 9,600 bits per second (9.6 Kbps). There are 16 X.25
lines at rates of 56 or 64 Kbps (80 percent utilization) using a
multiple mix of packet lengths, with a maximum of 128
bytes/packet for line handling. Packet rates are 64
packets/second. The remaining lines may be at any allowable
speed(s).
During average busy hours, call setup and 
takedown delay is
limited to 
53 seconds for 99 percent of all calls. Call setup and
takedown rates (Table 12.8) translate an average 15 CCS per
port for a 2-minute call holding time.
Table 12.8Call Setup and 
Takedown Rates
PEAK CALL SETUP ANDMAXIMUM NO. OF
NUMBER OF DATA
TAKEDOWN RATECALLS DURING A BUSY
PORTS
(CALLS/SECOND)MINUTE RATE
(CALLS/MINUTE) ,
64315
120424SVR 5210
8187s-201 
						

							TL-130500-1001
Data Traffic Per
12.9.8 The following equations can be used to calculate thePort Calculationnumber of ports which can be configured per LPB. These
equations assume a 25 percent overhead for internal control
traffic. Asynchronous and synchronous devices of varying
speeds may be mixed on each LPB. The curves in Figure 12.8
were calculated by using the following formulas:
1. The number of asynchronous devices per LPB is calculated as
follows:
s-202
12,OOOMP8 Bytes
N(a) =
SetX10 Bits XByteMP
1.25 xR Bits
Set ‘”
768,000N(a) =
RX”
where:
NWR
U= number of asynchronous devices per LPB
= baud rate of devices
= average bandwidth utilization by asynchronous
devices
and:Maximum N(a) is 64 ports
8187SVR 5210