University of Houston Networking 102 Dr Fred Zellner [e-mail]

1
University of HoustonNetworking 102Dr Fred
Zellnerfzellner_at_uh.edu

• Clocking

2
Clocking in the Digital Network

• Stratum Levels
• Clocking Distribution
• Single Point versus Multiple Points
• Timing Differences
• Correcting Timing Errors
• Buffers

3
Clocking

• The most important aspect in digital networking
  is clocking. The clock that quantifies the analog
  signal must be the same clock that reconstructs
  the signal at the other end. The only way to
  assure that the clock is accurate within the
  entire digital network is to have only one clock.
  

4
Clocking

• This is the way the first digital network was
  designed and setup in America. ATT set up a
  atomic clock which sent its time signals to each
  of the switching offices by means of the internal
  data network.

5
Clocking

• The different levels of clocking accuracy are
  called Stratum levels. This source of clock was
  called Stratum-1. Telephone companies distribute
  the analog clock frequency to the first working
  level which was called Stratum-2. This clock
  source is still very accurate but is starting to
  show signs of a measurable difference.

6
Clocking

• 1) Clock Recovery
• 2) Isochronous Clocking
• 3) Ones Density
• 4) The Phase Locked Loop (PLL)
• 5) Digital Data Services or 56 Kbps Data

7
Clocking

• Getting from a Stratum Level-2 to a Level-3
  telephone company office is achieved primarily
  over the digital network. The accuracy of a
  Stratum Level-3 clock relies on averaging time
  from a number of incoming digital lines, which
  results in a new nodal clock. Telephone company
  offices with Digital Cross-Connect Systems (DCCS)
  and DDS Hub offices use this Stratum level-3
  nodal clock. It is the highest level of timing
  accuracy accessible to end-user facilities.
• .

8
Clocking

• This now brings us down to the clock internal to
  the end-users terminal. Since the terminal
  relies on phase locking (PLL) this frequency from
  an internal crystal , the results are going to be
  less accurate than the clock coming to your
  location. In order to connect to the public
  network, you must have receive timing on
  transmit. This means you must use the clock
  coming in and the clock generator for the signal
  going out.

9
Clocking Distribution
Master Clock
Regional Message Switch
Network Timing Source
Digital Cross-Connect
Toll Office Switch
Digital Office Timing
End Office Switch
Digital Channel Bank
End User Terminal Equipment - Customer Owned
Switch
10
Clocking Distribution

• Telephone Company Distribution
• How many level-1 clocks exist in a particular
  telephone company network depends on their basic
  needs. Trans-Canada Telephone selected two master
  clocks--One at Ottawa and the other in Calgary.
  ATT had one master Clock at Hillsboro.

11
Clocking Distribution

• Telephone Company Distribution
• Other United States telephone companies have
  opted for independence from ATTs Reference
  frequency and have installed their own master
  clocks. Both Sprint and MCI have their own
  clocking schemes and employ multiple high level
  clock sources.

12
Telephone Company Distribution

• Many private companies have their own clock to
  time their networks because multiple carriers
  used or the number of different types of
  equipment at the customer site. As more Stratum
  Level-1 sources spring up across the country,
  look for a slight increase in bit slips between
  networks.

13
Clocking Distribution

• Satellite
• Telephone companies and some other operators of
  Large area networks use timing sources based on
  Stratum-1 clock sources located in satellites.
  Time can be broadcast to all of the network
  office facilities at the same time, there-by
  assuring the same clock time at all locations.

14
Typical Clocking Distribution
Clock 1
Clock 3
Clock 2
15
Clocking Differences

• Integrated digital networks are timing clock
  oriented. Clocks open and close logic gates to
  expected data pulses. When two clock sources
  appear in the same network, confusion abounds and
  errors ensue. Staying within one network only
  lessons the problems encountered. Your main
  concern is to reduce timing difference to a
  minimum, and thereby limit the number of bit
  slips in the network.

16
Clocking Differences

• There are several ways to reduce the clock
  differences. One method is adequate buffering
  between equipment and facilities. Another is
  slave-timing analog portions of your network to
  the main clock source.

17
Clocking Differences

• BIT SLIPS
• If you have two clocks in the same network, the
  difference between them will eventually result in
  a bit slip. A bit-slip results from one one clock
  source being faster or slower than the other. How
  often a bit-slip occurs depends on the timing
  difference. Digitized voice circuits go unscathed
  from even very high rates of bit-slips.

18
Clocking Differences

• Analog data circuits without built-in error
  correction may experience an error. Never the
  less, digital data circuits feel even the lowest
  number of bit-slips. Bit-slip can be either an
  extra bit inserted in the data stream or the
  omission of an expected bit.

19
Clocking Differences

• BIT STUFFING
• Reducing the clock difference between telephone
  company offices is a controlled timing
  difference. Telephone multiplexes purposely adds
  bits into the composite stream to adjust for the
  differences. Bit stuffing is a better name for
  this process. Stuff bits always appear in the
  aggregate data stream. If the transmitting end
  senses a need for a stuff bit, it notifies the
  other end to use one of the bits.

20
Clocking Distribution - Error
Correction

• If data comes in too slow
  the buffer runs runs low and the
  pointer will reset at the
  middle. Data comes in either too fast or
  too slow. If data comes
  in too fast the buffer
  overflows the pointer will reset at
  the middle

21
Clocking Differences

• BUFFERING
• Buffering provides a way to reduce the number of
  slips between two clock sources. It doesnt
  eliminate slips, but instead delays when they
  will happen. The buffer, after a bit slip goes
  back to the staring position. When the read clock
  is faster than the write clock, there comes a
  time when the buffer doesnt have any space in it
  or any bits to give. This results in a bit gap in
  the data output. Bit gaps also become bit slips.

22
Clocking Differences

• BUFFERING
• If the Write Clock is faster than the Read clock
  the buffer will fill up, resulting in the buffer
  emptying, causing a bit slip. Buffers have a
  major drawback adding delay to the network.
  Increasing the buffers size to extend the
  interval between bit slips, increasing the delay
  within the system.

23
Clocking Differences

• COMPOSITE FREQUENCY SUPPLIES
• Network offices have primary frequency generators
  that phase lock to an in coming master frequency
  source. The primary operating frequency in any
  telephone network is 4kHz. Multiples of 4kHz run
  anything from the individual channel to the very
  highest facility carrier system. A composite
  clock supply furnishes both 8kHz and 64kHz.
  Timing the Digital Signal Equipment (1 digital
  voice call) needs both the 8kHz and the 64Kbps
  clock

24
Clocking Differences

• CLOCK RECOVERY
• At each point in the network the clock is
  recovered by a Phase Locked Loop circuit. Every
  device that has a connection to the network must
  conform to loop timing on Transmit. What this
  means is that every data stream coming into a
  customer interface must trigger a PLL circuit.
  This clock generated by this PPL will loop the
  input clock back to the transmit side of the
  interface. This is done to make sure the timing
  accuracy level of the clock coming into the
  customer equipment is the same clock that will go
  back into the network facility.

25
Network Clocking
Network Clocking
Repeater
Repeater
Repeater
Repeater
T1 Multiplexer
T1 Multiplexer
CPE Equipment
CPE Equipment
CPE Equipment
Regenerator Phase-Locked Loop
26
Clocking

• ISOCHRONOUS
• Isochronous clocking is the type of clocking used
  in the digital telephony network. The PLL
  produces a constant clock signal signal using the
  incoming bits as triggers to maintain the
  accuracy of the circuit. You can view this
  circuit as kind of a flywheel. Once the flywheel
  gets up to speed it will maintain the same RPM
  with only a small amount of energy to keep it at
  the same speed. The energy used in the PLL is the
  incoming pulse.

27
Clocking

• ISOCHRONOUS
• If the incoming signal comes in before the
  circuit expects it the PLL speeds up a little as
  to adjust the clock output. The opposite occurs
  when the signal comes in after it is expected. It
  is necessary to make sure the network has a
  certain ones density to keep all of the PLL
  circuits in the network connection at the optimum
  timing accuracy.

28
Thank You !