Link Layer Flashcards
Describe the purpose of the Link Layer and some popular hardware implementations.
The link layer carries IP packets and Address Resolution Protocol (ARP) packets.
It is primarily responsible for transferring data from the source machine’s network layer to the destination machine’s network layer.
Ethernet, ADSL, PPP, VDSL, ATM, Wireless protocols, and Cellular protocols are popular hardware implementations of the link layer.
What are the types of services the link layer provides?
There are three types of services the link layer provides:
1. Unacknowledged connectionless service
- Acknowledged connectionless service
- Acknowledged connection-oriented service
Describe the unacknowledged connectionless service in the Link Layer.
In an unacknowledged connectionless service, no connection is required, and packets are sent without acknowledgement. This can be problematic for the receiver if some packets are dropped en route, and the receiver may never fully receive the data.
Additionally, packets are not necessarily in order, and the receiver has to figure out the order.
Unacknowledged connectionless service is primarily used in LAN by the link layer. It is also valuable for real-time data services, where late data is worse than no data (e.g., streaming a video).
Describe the acknowledged connectionless service in the Link Layer.
In an acknowledged connectionless service, no connection is required and each packet sent is acknowledged.
This adds a level of reliability compared to its unacknowledged counterpart. However, this adds considerable overhead to the link layer as each packet is to be acknowledged.
Additionally, if an acknowledgement packet is dropped en route, it can cause the sender to send the same data again; and the receiver to receive the same data again.
It is helpful for broad channel usage, such as wireless.
Describe the acknowledged connection-oriented service in the Link Layer.
An acknowledged connection-oriented service establishes a connection between two machines before there is any data transfer.
Each packet is sequentially numbered and sent precisely once. The packets are guaranteed to be sent and received.
This provides network layer processes with a type of bit stream.
There are 3 phases of this service:
1. The connection is established between two machines, and variables are initialised to track which frames have been sent.
2. One or more frames are transmitted
3. The connection is released, freeing all variables, buffers, and resources needed to maintain the connection.
Describe ADSL and explain why it is asymmetric.
Asymmetric Digital Subscriber Line (ADSL) was a popular method of network connectivity delivery to the home. Many ISPs use ADSL as a layer to support ATM.
ADSL takes advantage of the ability of a telephone wire to support multiple frequencies. It uses many blocks of frequencies simultaneously – broadband – that avoid areas of the spectrum that have interference.
It is asymmetric in that it divides the available bandwidth unequally into 24Mb/s downstream and 2Mb/s upstream. So there isn’t a symmetric data rate on the copper wire.
Describe the ADSL Frame and the kinds of data it contains.
Data in an ADSL Frame is transmitted in a ‘superframe’ containing 69 ADSL frames. Frame 69 is for synchronisation – rest contain data.
An ADSL frame contains two kinds of data: fast data and interleaved data.
The fast section is used for time-sensitive data such as audio and can contain forward error correction bits.
The interleaved section has purely user data. It is called interleaved because the data bits are mixed into an interleaved order which is intended to help against errors.
Describe the Last Mile Problem and potential solutions for it.
The last mile problem is the issue of bridging the gap between the local telephone exchange and the final user.
There are many potential solutions being tried:
1. Fibre to the Home connects each customer with new fibre. This gives many megabits of bandwidth but is typically too expensive as it requires digging up every street in the country
2. Fibre to the Kerb/Cabinet cables up every street and a short copper wire connecting from the street cabinet to the home. It is cheaper but still a lot of fibre must be laid
Describe the SLIP protocol.
Serial Line Internet Protocol (SLIP) is an early protocol used on models to encapsulate IP traffic over serial (telephone) lines.
It is a point-to-point protocol, meaning it links just two machines to each other – this was the normal requirement in early dial-up systems.
Describe how the SLIP Frame encapsulates data.
A SLIP frame encapsulates data using terminating bytes padded on both sides.
There is no frame size limit for a SLIP frame. But 296 bytes was common (40 bytes of TCP/IP headers plus 256 bytes of data).
What is byte stuffing?
Byte stuffing is the technique of adding an escape character byte (ESC) to user data if it contained an padding character.
The sender’s data is inserted with ESC bytes wherever it contains a padding character. The receiver then finds the ESC bytes and removes them before passing to the above layers.
What are some of the problems with SLIP?
SLIP only supports IP – there is no type field in its frame. Which means that it is not suitable for transmitting different protocols over the network.
The ends must have pre-agreed IP addresses – there is no mechanism for agreeing addresses. When you connect your machine to the network, you don’t really know what kind of IP you have.
There is no checksum present: early telephone lines were very noise and created data corruption. There was nothing you could do about it.
There is no authentication present: there is no way of checking who is connecting.
Describe the PPP protocol.
The Point-to-Point protocol is an evolution of SLIP to fix its problems. It was originally designed to be used as a link layer over telephone lines.
PPP has three parts:
1. A framing layout for packets
2. A link control protocol (LCP) for managing, configuring, and authenticating links
3. A set of network control protocols (NCP) to manage network layer specific otpions. This includes features like header compression, authentication, and frame size.
You could have a faster connection by multiplexing telephone lines and using Multilink PPP.
Describe PPPoE.
PPP over Ethernet is PPP layered over Ethernet. This provides the ability to manage access control, billing and level of service – things that are important to ISPs.
Fibre-to-the-cabinet (FTTC) products use PPPoE over VDSL to pass authentication to the ISP.
PPPoE is a prime example of why we need layering - PPP is wrapped around Ethernet.
Describe the PPP Frame.
A PPP frame has the following fields:
- Frame delimiters: padded at the start and end
- Address field
- Control field
- Protocol field: tells us what the next layer is
- Cyclic redundancy check (CRC)
A PPP frame holds up to 1500 bytes of data.
