Lower Layers Part 2 Flashcards
What are the three types of networks classified by area?
LAN (Local Area Network), MAN (Metropolitan Area Network), WAN (Wide Area Network)
Personal Area Networks (PANs) are also sometimes included.
What is the primary distinction between LANs and WANs in terms of technology?
LANs typically use simpler technologies and protocols, while WANs require more complex planning and structured addressing.
What does CSMA/CD stand for?
Carrier Sense Multiple Access Collision Detection
What was the main purpose of early LANs?
To connect terminals to timeshare computers
What is one of the main applications of WANs?
File transfer, job transfer, remote login
True or False: Ethernet was originally developed at Xerox in the 1970s.
True
Fill in the blank: The maximum frame size for Ethernet is _______ bytes.
1500
What is the significance of the 0x55 preamble in Ethernet frames?
It allows receivers to synchronize with the data transmission.
What technology is known for connecting computers within data centers?
Vendor Networks
What is the typical maximum length of an Ethernet segment without amplification?
500 meters
What protocol suite did ARPANET originally use before TCP/IP?
NCP (Network Control Protocol)
What are the two main types of Ethernet cables mentioned?
Coaxial cable and twisted pair cable
What happens when a collision is detected in a CSMA/CD network?
The network is jammed with a specific pattern to inform all stations of the collision.
What is the minimum frame size for Ethernet?
64 bytes
True or False: The function of the VLAN tag in an Ethernet frame is mandatory.
False
What does ATM stand for in networking technology?
Asynchronous Transfer Mode
What is the purpose of the inter-packet gap in Ethernet?
To allow time for the network to recover before the next packet is sent.
What is the main challenge of WAN technology historically?
Slow speed and unreliability
What is the maximum diameter for an Ethernet network determined by?
The maximum length of the transmission medium and the permissible number of repeaters.
What is the function of a checksum in Ethernet frames?
To verify the integrity of the data within the packet.
What does the term ‘collision domain’ refer to?
The set of all stations whose packets might mutually collide.
Fill in the blank: Ethernet originally operated at _____ Mbps.
3
True or False: The Ethernet protocol allows multiple stations to transmit simultaneously without any control.
False
What are the different BaseT standards for Ethernet?
10BaseT, 100BaseT, 1000BaseT
These standards represent different speeds and technologies for Ethernet networking.
What is the maximum length for twisted pair cables in modern Ethernet?
95m
This applies when using RJ45 connectors to a hub.
What type of cabling is required for faster Ethernet speeds?
Category 5
This cabling has different twists in the pairs to support higher data rates.
What is the function of a repeater in networking?
An analogue amplifier that sees collisions on both sides
Repeaters regenerate signals but do not prevent collisions.
What distinguishes a ‘dumb’ bridge from a ‘learning’ bridge?
‘Dumb’ bridges propagate all frames, while ‘learning’ bridges only send frames that belong on the other side
This reduces unnecessary traffic in the network.
True or False: Ether hubs are considered bridges.
False
Ether hubs are repeaters and do not manage collisions like bridges do.
What is the main advantage of full duplex in Ethernet?
Traffic goes in and out without colliding
Each direction operates as a separate collision domain.
What are cut-through switches designed to do?
Transmit frames immediately after reading the header
This reduces latency but can propagate broken frames.
What is a key characteristic of ATM (Asynchronous Transfer Mode)?
It breaks data into a stream of 48-byte packets
This was a compromise due to political reasons among different stakeholders.
What is the purpose of traffic shaping in ATM?
To manage bandwidth and ensure quality of service
It allows for different profiles like Constant Bit Rate and Variable Bit Rate.
What is the maximum speed of ATM25?
25Mbps
This was commonly used in ADSL modems during the late 1990s.
What networking topology does IBM Token Ring use?
Star topology
This topology was popular for its robustness.
Fill in the blank: A __________ is a set of learning bridges in a box.
switch
Switches manage traffic efficiently in a network.
What is the latency introduced by conservative switches?
Additional latency compared to a straight piece of wire
This is due to the need to confirm checksums before transmission.
What is the function of a virtual circuit in ATM?
To establish a dedicated path for packets
This simplifies error recovery and maintains packet order.
What problem is associated with the original idea for UK ADSL broadband using switched virtual circuits?
ATM switches couldn’t handle the volume of circuit establishment
This led to the use of permanent virtual circuits instead.
What is the main disadvantage of the ATM’s 48-byte packet structure?
It is appalling for performance and reliability
This structure complicates the mixing of voice and data.
True or False: Token Ring networks guarantee bounded latency.
True
However, in practice, achieving this is complicated.
What technology is used in CDDI?
FDDI over copper
CDDI can survive multiple failures but was expensive.
What is the purpose of slotted rings in networking?
To circulate empty data frames instead of tokens
This concept was inspired by the conveyor belt in a Sushi restaurant.
What does the term ‘echo cancellation’ refer to?
Filtering out echoes in telephony for clear audio
This is critical for maintaining call quality.
What are the four types of ATM traffic profiles?
- Constant Bit Rate
- Available Bit Rate
- Variable Bit Rate
- Unspecified Bit Rate
These profiles help manage bandwidth and quality of service.
What is a Constant Bit Rate (CBR)?
Guarantees a fixed-speed point-to-point link
CBR is as close to a piece of wire as you can get.
What does Variable Bit Rate (VBR) allow?
Allows bursting to higher speeds while preserving guarantees about timing
What is Available Bit Rate (ABR)?
Guarantees a minimum performance but potentially delays packets that exceed it
ABR is very complicated.
What does Unspecified Bit Rate (UBR) represent?
Best efforts for rate, delay, and reliability
What do all traffic profiles guarantee that IP does not?
Ordering
What is traffic shaping?
Shaping data into a particular profile by clocking it into large buffers as it arrives
What happens when the average input rate exceeds the average departure rate in traffic shaping?
Packets are dropped
What is Random Early Drop?
A strategy to randomly drop packets with increasing probability as the buffer fills
What is the purpose of traffic policing?
Checks whether passing traffic is within parameters for the virtual circuit and discards anything outside
What is Multi-Protocol Label Switching (MPLS)?
A technology that can be thought of as ATM with larger packets
What is the benefit of using Wave Division Multiplexing (WDM)?
Transmits multiple streams down a single fibre using different colour light
What is the difference between Dense and Coarse Wavelength Division Multiplexing?
Coarse: 20nm difference; Dense: originally 0.8nm difference, now can be 0.4nm or 0.2nm
What is the typical use case for Ethernet in 2023?
Short range: ether over copper; Medium range: ether over multimode fibre; Long range: ether over WDM
What does Ethernet switching solve?
Many problems of Ethernet including data hogging by machines and unpredictable performance
What is understeer in driving terms?
When the driver hits the barrier with the front of the car
What does queuing in networking refer to?
The management of data packets as they wait for processing in a network
What can happen if a switch does not have ingress queuing of at least 8K?
The request is repeated until the PC gives up
What is a special case of queuing in networking?
Blocking
What are the characteristics of core switches?
Non-blocking, sufficient capacity to switch all ports at wire speed with arbitrary packet sizes
What is the typical buffering strategy at the edge of a network?
Deeper per-port buffers and sometimes global buffers shared by all ports
What is the takeaway regarding Ethernet switching in modern networks?
Robust under load with substantial excess performance
True or False: Debugging in network capacity issues is straightforward.
False
Fill in the blank: Traffic policing is often called ______.
hard drop policing
What job did early LANs perform in addition to connecting terminals?
Early LANs were also used for tasks that WANs performed but achieved them faster.
What is UUCP and why is it mentioned?
UUCP (Unix-to-Unix Copy Protocol) is a system for transferring files, executing commands, and sending email between Unix systems. It was widely used in the early days of networking before the internet became dominant.
It is mentioned humorously here as a relic of the past, with the note that it hasn’t been “used in the wild” for around 20 years.
What was a turning point in the past for needing faster networks?
When personal workstations were introduced.
Why is the 0x55 pattern used in Ethernet communication?
The 0x55 pattern is used to help receivers lock to the current speed of the transmitter. This is crucial because Ethernet has no central clock, and devices must synchronize their data clocking independently.
What does the note “Network time protocol is standardized now” imply about Ethernet clocking?
The note suggests that while Ethernet still relies on decentralized clocking (like using 0x55 patterns), standardized time protocols (e.g., NTP) can help align clocks across devices for better synchronization.
What is the process for recovering from a collision in Ethernet communication?
On the first attempt, choose a “random” number ( k ) from {0,1} and delay ( k \times 512 ) bit periods before retrying. For subsequent attempts, choose ( k ) from {0…( 2^n )} and delay ( k \times 512 ) bit periods, giving up after 10 attempts.
Does the “random number” used in collision recovery need to be truly random?
No, it doesn’t have to be truly random, just different from other stations to generate non-identical sequences of delays.
What is the maximum number of attempts allowed for collision recovery?
The maximum number of attempts is 10, after which the system gives up.
What does a Dumb Bridge do with traffic between hosts?
A Dumb Bridge forwards all unbroken frames, making traffic visible across both sides.
When using a Dumb Bridge, can traffic from Host 1 to Host 2 collide with traffic from Host 4 to Host 5 in a Dumb Bridge setup?
No, traffic between these groups does not collide.
What is the main feature of a Filtering Bridge?
A Filtering Bridge forwards unbroken frames except for local traffic.
Why is a Filtering Bridge better for security and capacity?
It limits visibility and reduces unnecessary traffic forwarding.
What happens to packets with unknown destinations in an Ethernet Switch?
Packets are sent out of all ports unless the traffic is for a specific connected device.
What is an Ethernet Switch?
An Ethernet Switch is a set of learning bridges in a box, with each interface having its own collision domain.
How do large buffers in Ethernet Switches handle congestion?
Buffers assume high data traffic will last only briefly, managing temporary bursts.
What is the backplane in an Ethernet Switch?
It allows traffic to be sent only when the destination is known or needs to be located.
Is the Ethernet Switch backplane slow or fast? What does it’s speed determine?
It is very fast. It’s speed determines the performance of the entire switch.
What is the difference between conservative and aggressive cut-through switches?
Conservative switches wait for the entire frame and verify the checksum, while aggressive switches start transmitting after reading the header.
What is the latency issue with conservative switches?
They introduce latency by waiting for the full packet, equivalent to several kilometers of copper transmission.
What is the benefit of aggressive cut-through switches?
They minimize latency by transmitting after reading only 160–192 bits of the header.
What is a potential problem with aggressive cut-through switches?
They might propagate broken frames as they do not verify checksums.
What was argued about Ethernet despite limited evidence?
It was argued to behave badly under high load.
What is a theoretical advantage of token-based systems for a network?
They offer bounded latency, as the token circulates in a predictable time based on the number of stations and packet periods.
What are practical challenges of token-based systems for a network?
They are complex to implement correctly due to issues like token loss, creation, and station failure.
What is FDDI Fibre, and where is it still in use?
FDDI Fibre is a dual-ring network offering 100Mbps, known for robustness, and is still used in interconnects and data centres as a backup.
What is a Slotted Ring network?
A Slotted Ring network requires a minimum cable length to maintain empty packets and often includes coiled cables under the floor.
Why was the Slotted Ring network popular in UK universities?
It was cheap, easy to build, and had driver support for Unix variants, though it saw little traction elsewhere.
What happens in a Slotted Ring network if a computer crashes after sending a packet that isn’t received?
A frame will remain occupied until a clean-up algorithm is used to clear it.
What was the issue with creating and destroying circuits in switched virtual circuits (SVC)?
Creating and destroying circuits was very “expensive”, making it impractical for frequent dynamic use.
What is traffic shaping in networking?
Traffic shaping smooths incoming data into a buffer, controlling output rate to prevent congestion and delay.
How does a buffer in traffic shaping work?
A buffer queues data like passengers on a crowded train, holding it until the network can transmit it at a controlled rate.
What is a crude solution for handling excessive traffic in networking?
discard packets to prevent buffer overflow.
What is Random Early Drop (RED)?
RED randomly drops packets as buffers fill up to signal senders to slow down, avoiding complete congestion.
What happens after packet loss in transport connections?
Packet loss triggers a timeout and retransmission, slowing down the connection over time.
What is traffic policing in networking?
Traffic policing, or “hard drop” policing, enforces parameters for virtual circuits and discards excess traffic.
When is traffic policing useful in networking?
Traffic policing is useful for managing mixed data types like voice, file transfer, and video conferencing.
What is Wave Division Multiplexing (WDM)?
WDM uses different colours of light to transmit multiple data streams on a single fibre for long-distance communication.
What is the key takeaway regarding Ethernet switching and robustness?
Ethernet switching is robust under load, with 10GigE interconnects offering substantial performance headroom.
What makes debugging in Ethernet switching environments challenging?
Performance issues, once found, can be dramatic and hard to isolate.
Why might over-engineering in networking be preferred?
It can be cheaper and more effective than trying to “right-size” resources for network demands.
What caution is advised regarding buffering in networks?
More buffering is better, especially at the edge of the network, to handle load.