bitmindframes

Standards and Implementation:

The first LAN ever was Ethernet.
Robert Metcalfe and coworkers at Xerox in mid 70s.
The first Ethernet standard published 1980 by DigitalEquipCorp, Intel, Xerox (DIX).
Metcalfe wanted it to be a shared open standard.
1985 IEEE published standards for LANs.
For compatibility with OSI, small mods made to Ethernet standards with 802.3.

Ethernet –L1 and L2:

Ethernet implemented in MAC sublayer and L1 only.

LLC:

The use of these sublayers contributes significantly to compatibility between diverse end devices.
802.2 describes LLC sublayer functions.
802.3 describes the Media Access Control (MAC) sublayer and L1 functions.
LLC – software (drivers) – independent physical equipment.

MAC (Media Access Control):

Implemented by hardware (NIC)
Primary responsibilities:
Data Encapsulation – framing, addressing, error detection.
Media Access Control – does what it says (all in the name, how).

Logical Topology:

Ethernet’s underlying topology is the multi-access bus.
All nodes share the medium.
All the nodes receive all the frames.
Each node checks every frame and asks, “is it for me?”

Physical Implementations of the Ethernet:

Almost all Internet traffic starts and stops with Ethernet.
Ethernet has adapted to new technologies for 30 years.
Why the success?
Simplicity and ease of maintenance.
Ability to incorporate new technologies.
Reliability.
Low cost of installation and upgrade.
Gb Ethernet is moving into MAN and WAN standards.
The Ethernet frame is consistent across all implementations.

Historic Ethernet:

Ethernet concepts based loosely on Alohanet (1970).
It was a digital radio network that used a shared radio frequency between the Hawaiian Islands.
CSMA/CD
The first versions used coax in a physical bus topology (think net, vampire taps).
Each PC was directly connected to the backbone.
Thicknet (10BASE5) and Thinnet (10BASE2).
10BASE5 (10Mbps, BASEband transmission, 500m without repeater).
10BASE2 (10Mbps, BASEband transmission, 185m without repeater).
BASE(band)– Digital.
Broad(band) – Analog.
10BASE5 – thick coax – up to 500m without repeater.
10BASE2 – thin coax – up to 185m.
Coax was replaced by early categories of UTP (CAT3).
Easier to work with, lightweight, and less cost.
Physical topology changed to a star, using hubs.
Hubs concentrate connections to one point (the hub).
Hubs increased reliability – did not solve collisions.

Ethernet Collision Management:

Legacy Ethernet (classic).
Shared media – always half-duplex.
More devices, more collisions –handled by CSMA/CD.
Current Ethernet (Switched Ethernet).
A significant development when switches replaced hubs.
Around the same time 100BASE-TX was introduced.
10BASE-T = 10Mbps twisted pair 100m
100BASE-TX
Switches send frames to the correct port – not every device.

Moving to 1Gb and Beyound:

Modern applications tax even robust networks (eg. VoIP/Video, etc.).
GbE (Gigabit Ethernet) provides bandwidth of 1000Mbps.
Significant increase in network performance.
Some equipment and cabling may be capable of working at the higher speeds with minimal upgrading. Reduces Total cost of Ownership (TCO).
Ethernet Beyond the LAN.
The increased distances enabled by fiber has blurred the distinction between LANs and WANs.
Ethernet – across a city – MAN (Metropolitan Area Network).

The Frame – Encapsulating the Packet:

2 types: 802.3 (original) and revised 802.3 (Ethernet).
Ethernet Frame size – 64 to 1518 Bytes ********* need to know *********
preamble and SFD (start frame delimitor), are not included in the size.
1998 ? 802.3ac extended max. to 1522 for VLANs.
If a frame < 64 or > 1518 the frame is dropped.

Ethernet 802.3 Frame Fields:

Preamble – 7 bytes + SFD – 1 byte
0101010 ….. 010111.
Both are used to sync devices. “Get ready for a frame”..
Destination MAC Address – 6 bytes (48bits).
Source MAC Address – 6 bytes (48bits).
Length/Type – 2 bytes – length of data OR type of data.
Only length in early version and only Type in DIX ver.
Combined in a later versions.
If greater then or equal to 0×0600 or 153610, then contents = type.
Data and Pad – 46 to 1500 bytes
Contains L3 PDU. If data is too small (not 46 bytes), pad to min size.
FCS – 4 bytes, uses CRC to detect errors.

Ethernet II Frame Field:

Eithernet II frames have 2 slight differences from 802.3.
Preamble – 8 bytes.
Still has the identical SFD, but rolled into this field.
0101010….01011
Basically includes the SFD.
Type – only shows protocol, NOT length.
All other fields remain the same.
Ethernet II – used for TCP/IP traffic.
802.3 – used for all other traffic.

The Ethernet MAC Address:

A unique ID for source and destination.
All Ethernet versions use same addressing.
48-bit value expressed as 12 hexadecimal digits.
Organizational Unique ID (OUI)
MAC Address Structure.
Vendors register with IEEE ? 3-byte OUI.
MAC = OUI + unique serial number = 6 bytes.
ADA BIA (Burned In Address) – burned into ROM on the NIC.
Cannot be changed by software (hardware).
Addresses can be formatted 3 ways.
00-05-9A-3C-78-00.
00:05:9A:3C:78:00.
0005.9A#C.7800.
First 6 hex digits are manufacturer specific.

Viewing the MAC:

ipconfig / all, windows
iwconfig /all windows wireless.
ifconfig, LINIX.
Also, via GUI in XP on support tab of Local Area Conn.

Another Layer of Addressing:

Data Link Layer.
A MAC address is for local media only.
Non-hierarchical – no meaning outside local network.
Carries a packet on local media across each segment.
Logical addressing (IPv4) enables the packet to be forwarded toward its destination.

Ethernet Unicast:

A unicast MAC is the unique address used to ID a node.
MAC actually delivers to the local host.

Ethernet Broadcasts:

IP broadcasts have all 1s in the host portion.
Needs a corresponding MAC address in Ethernet.
MAC broadcast = 48 ones = FF-FF-FF-FF-FF-FF, (remember only 6 bytes).
Can only be used as a destination.

Ethernet Multicast:

Devices in a multicast group are assigned a multicast IP.
Only used as the destination.
The multicast IP requires a corresponding multicast MAC.
A special value that begins with 24 bits, 01-00-5E hex + 1 bit, 0 binary + the last 23 bits of the IP address.

Media Access control in Ethernet:

Collisions are the cost for Ethernet to get the low overhead.
Ethernet uses CSMA/CD to handle collisions.
Carrier Sense:
- All devices that need to send must listen first.
If another signal detected, wait.
When no traffic a device will transmit its message.
While transmitting, continue to listen for collisions.
Multi-access:
A collision is detected as an increase in voltage on the media.
Once detected, every device transmitting will continue to transmit to ensure that all devices on the network detect the collision.
Jam Signal and Random Backoff:
Once detected, every Tx device in the collision sends a jamming signal, 32 bits (4 bytes).
This notifies other devices to invoke a backoff algorithm.
Backoff algorithm causes all devices to stop for a random time.
After that the device goes back to “listening before sending”.
A random backoff period ensures that the devices that were involved in the collision do not try to send their traffic again at the same time.
This could mean that a 3rd device may transmit before either of the 2 involved.
9.4.2 Animation (good to view).

Hubs:

L1 device.
Multiport repeater.
“Shared” media.
Do not filter data in any way.
The devices that access a hub or hubs = collision domain.
AKA a network segment.
Hubs and repeaters increase the size of the collision domain.
Interconnected hubs form an extended star topology.
An increased number of collisions reduces efficiency use of the bandwidth.

Ethernet Timing:

Faster Ethernet makes collision management more complex.
Latency – time taken to propagate (travel) down the cable (propagation delay).
Each hub adds latency forwarding bits from port to port (each device adds a bit of a delay).
Accumulated delay increases the likelihood of collisions, (one end of the wire does not hear others until delay is reached and may start to TX).
A listening node might Tx while the hub is processing.

Timing and synchronization:

In half-duplex, devices Tx 64 bits of synchronization information (preamble).
And then Tx the complete frame.
10 Mbps Ethernet is asynchronous – in this context means that each receiving device will use the 8 bytes of timing information to sync the rcv circuit to the incoming data and then discard the 8 bytes.
100 Mb and higher are synchronous – in this context means that the timing info is not required. However, for compatibility reasons, the Preamble and Start Frame Delimiter (SFD) fields are still present.

Bit Time:

The time required for a bit to be placed/sensed on media.

10 Mbps > 100ns.
100 Mbps > 10ns.
1Gbps > 1ns.
10 Gbps > 0.1ns.

As a rough estimate, 20.3 cietimeters (8 inches) per nanosecond is often used for calculating the propagation delay on a UTP cable.
The result is that for 100 meters of UTP cable, it takes just under 5 bit times for a 10BASE-T signal to travel the length of the cable.
For CSMA/CD to work, a Tx device must sense a collision before it has finished a minimum-sized frame (64 Bytes or 512 bits).
At 100 Mbps, the timing barely accommodates 100 meters.
At 1000 Mbps nearly an entire min. frame (64 Bytes), would be transmitted before the first bit reached the end.
For this reason ,half-duplex mode not permitted in 10-Gb Ethernet.

10Mbps = 1
——————-
10,000,000

Slot Time:

Ensures that if a collision occurs, it will be detected.
Below 1 Gb, a transmission must last at least one slot time.
Allows Tx to send a min sized frame and still detect collision at the last possible minute.
10Mb and 100Mb = 512 bits (4096 for Gb).
Min. sized frame = 64 Bytes, if you take this and multiple it by 8 bits you get 512 bits. Therefore, there are 512 bits in 64 Bytes.
Slot time in legacy Ethernet is 51.2ms.
Trade-off between the need to reduce the impact of collisions and the need for reasonable network sizes.
Max network diameter = 2500 meters.
Slot time for 10Mbps and 100Mbps Ethernet 512 bit times, or 64 octets. Slot time for 1000Mbps Ethernet is 4096 bit times, or 512 octets.
The slot time ensures that if a collision is going to occur, it will be detected within the first 512 bits (5096 for Gigabit Ethernet) of the frame transmission.
The 512-bit slot time establishes the minimum size of an Ethernet frame as 64 bytes. Any frame < 64 Bytes = collision fragment or runt frame and is automatically discarded by the receiving stations.
Establishes a limit on the max size f a network’s segments. If a network gets too big, late collisions can occur.
Late collisions are considered a failure in the network because the collision is detected too late by a device during the frame transmission to be automatically handled by CSMA/CD to work
Is calculated assuming max cable lengths on the largest legal network architecture. All hardware propagation delay times are at the legal max and the 32-bit jam signal is used when collisions are detected.
Slot time is just greater then the time required to travel between the furthest points of the collision domain, collide at the last possible instant, and then have the collision fragments return to the sending station and be detected (see figure).
For the system to work properly, the first device must learn about the collision before it finishes sending the smallest legal frame size (64 bits).
To allow GbE in half-duplex, an extension field is added to small frames to keep the transmitter busy long enough for a collision fragment to return, (Cisco question).

5-4-3 rule:

Repeaters make Ethernet collision domains larger.
Legacy Ethernet specifies a maximum of:
5 Segments.
4 Repeaters.
3 Populated Segments.

Does not apply switched Ethernet.
Cisco used to teach as, 5-4-3-2-1
2-unpopulated segments.
1- collision domain.

Tags: study, wan, ip, cisco, interface, guides