Karteikarten
SCC 203 - Networking
Karten in diesem Stapel (200)
- 1What is circuit switching?A mode of data transmission which consists of 3 distinct phases
Call set-up
Communicate
Tear down - 2What is packet switching?A mode of data transmission in which a message is broken up and sent over a network as 'packets'
- 3Define protocolA protocol is a specification of all the possible events and actions within a system
- expected response to every action
- exact details of every message sent - 4What are the two types of addresses?Hierarchical - has prefixes that allow us to infer location from an address,
topological location meaning they're determined by internal network structure
Flat - consists of blocks of address which are allocated by vender. Given address no idea where device using it is - 5What are the two types of connection?Connection oriented - uses circuit switching
-resources held for connection end-to-end
-resources still tied up if connection is idle
-bandwidth is partitioned
Connection-less - uses packet switching - 6What is a simple packet split into and describe each part?Header - holds address or ID for path selection, which is used to determine how we get from sender -> receiver
Payload - holds application data - 7What are the two modes involved in packet switching and what do 'modes' determine?Datagram - apps aware of packet structure
Virtual Circuit - apps unaware of packet structure
Modes determine the style of communication and how we get from A-B - 8Describe the datagram mode?The datagram mode is where each datagram packet is considered a separate message
header -> sender & destination address
payload -> application data - 9Describe the virtual circuit mode?The virtual circuit mode emulates connection-orientation over connectionless network
header -> connection tag (sender & receiver)
payload -> application data - 10What are the two views and what do they consist of?Application view - this is the service provided to applications
Consists of;
Datagrams (connection-less)
Virtual circuits (emulates connection-oriented over connectionless
Physical view - how the underlying network operates
Consists of;
Circuit switching (connection-oriented)
Packet switching (connectionless) - 11How do you calculate the transmission delay?Length of packet / bits per second
- 12How do you calculate the propagation delayLength of link / propagation speed (based on material transmitting over (e.g. copper wire, fibre optics)
- 13What is a communication protocol?The specification of how data is transmitted, and when it is received
Allowed messages/expected replies
Ordering/timing of transmissions
Format of transmitted data - 14Why do we use layers and what do layers provide?We use layers to abstract out complexity and to provide a service to the layer above
- 15What are the layers in the OSi-ISO model?Application layer
Presentation layer
Session layer
Transport layer
Network layer
Data Link layer
Physical layer - 16Describe the physical layer and what is transmittedThe physical layer is the mechanical and electric specification
e.g. type of wires
Symbols are transmitted - 17Describe the data-link layer and what is transmittedThe three main functions of the data link layer are to deal with transmission errors, regulate the flow of data, and provide a well-defined interface to the network layer
Frames are transmitted - 18Describe the network layer and what is transmittedThe network layer consists of independent host-to-host transmission
- 19Describe the transport layer and what is transmittedThe transport layer is responsible for delivering data to the appropriate application process on the host computers
The transport layer has a user oriented global addressing scheme
Segments are transmitted - 20Describe the session layerThe session layer provides the mechanism for opening, closing and managing a session (connection) between end-user application processes
- 21Describe the presentation layerThe presentation layer is responsible for the following
Data encryption/decryption
Character/string conversion
Data compression - 22Describe the application layerThe application layer is an abstraction layer that specifies the shared protocols and interface methods used by hosts in a communications network
- 23What are the layers in the Internet model?Application layer
Transport layer
Network layer
Link layer
Physical layer - 24Describe what encapsulation is and the processEncapsulation adds control information as information is passed to layer below
Transmitting (encapsulating) - headers are added at each layer as the information is passed
Receiving (de-capsulating) - headers are stripped as the data is passed up the layers - 25What is the structure of a datagram packet header?Destination - 32 bits
Sender - 32 bits
Length - 16 bits - how many bytes packet is
Type - 16 bits - 26How does a 'store and forward' network work?Packets received
Stored locally
Forwarded to destination - 27How does the data flow in the OSI model?Start highest layer, work way down
Enter intermediate systems at lowest layer
Pass back to end systems lowest layer and work up - 28What is binary encoding and pros/cons?Pro's
+ Simple, cheap
+ Good utilization of the bandwidth (1 bit per Baud)
Cheap
- No "self-clocking" feature - 29What is Manchester encoding and pro/cons?Pros and Cons
+ good "self clocking" feature
- Only 50% of the binary coding bandwidth utilisation
- 30What is Differential Manchester Encoding and pros/cons?Manchester encoding but every 1 bit the sequence will switch
- 31What is a clock signal?A clock signal is a particular type of signal that oscillates between a high and a low state and is utilized like a metronome to coordinate actions of digital circuits.
- 32What is Frequency Division Multiplexing (FDM)?Frequency bandwidth is split between users
User allocated a frequency band
Used in for phoneline - 33What is Time Division Multiplexing (TDM)?User receives time slot, during this time user has full bandwidth
- 34What is synchronous TDM?Allocate in fixed slot per channel
sender can only send in allocated slot
Channel 1 - Channel 2
A1 B2
C1 C2
A1, B2, C1, C2 - 35What is Asynchronous TDM?Packets are processed according to arrival time
Slot - Slot - Slot - Slot
A1
B1
A1, B1 - 36Define framingFraming is the process of encapsulating a datagram into a frame, adding a header and trailer
MAC addresses are used in frame headers
Providing reliable delivery between adjacent nodes
Header - payload - trailer - 37What are some link layer services?Flow control - controls the pacing between adjacent sending and receiving nodes
Error detection - detects errors caused by signal attenuation, noise
Error correction - identifies and corrects bit error(s) without
resorting to retransmission
Half duplex - communication between 2 nodes can only happen one at once
Full duplex - communication between 2 nodes can happen simultaneously - 38How does parity checking work?If odd number of 1's = 1
even number of 1's = 0 - 39What is Medium Access Control (MAC)?Point-to-point connections, typically uses simple two-way handshake
- 40What are the two link types?Point-to-point - link between Ethernet switch and host
Broadcast - shared wire - 41What are the two sub-layers of the Data Link Layer in the OSI-ISO model?Logical Link Control (LLC) layer - presents interface to the user of the data link service
Medium Access Control (MAC) layer - ensures that data is transferred correctly between adjacent network nodes - 42What do Random Access MAC protocols specify and what is an example of one?How to detect collisions
How to recover from collisions
Slotted ALOHA
ALOHA - 43How does Slotted ALOHA work?If no collisions occur, next frame can be transmitted in the next slot
If collision occurs, retransmit in the next slot available, keep trying until success - 44What are the pros/cons of Slotted ALOHA?Pros;
Simple
Nodes can transmit at full rate of the channel
Only slots in nodes need to be in sync (highly decentralized)
Cons;
Collisions, wasting slots
Idle slots
Clock sync
Nodes may be able to detect collisions in less than time to transmit packet - 45How does pure ALOHA work?When frame first arrives, transmit it immediately
No sync required
the slides on this are wank - 46What is CSMA?Carrier Sense Multiple Access
Listens before transmitting
if channel sensed idle - transmit entire frame
if channel sensed busy - defer transmission - 47Why can collisions still occur in CSMA?Because the propagation delay means that two nodes may not hear each others transmission
Meaning entire packet transmission time is wasted - 48What is CSMA/CD?CSMA with collision detection
Colliding transmissions are aborted, reducing channel wastage
easier in wired LANs as oppose to wireless LANs - 49What are the two physical topologies of Ethernet?Bus - all nodes in same collision domain
Star - active switch in the center, each link runs a separate Ethernet protocol - 50Describe an Ethernet SwitchLink layer device
Stores and forwards Ethernet frames
Examines incoming frame's MAC address, forwards frame to one-or-more outgoing links when frame is to be forwarded on segment
Transparent - meaning host unaware of presence - 51How do 'Multiple Simultaneous Transmission' work when using a Switch?Each host connected to the switch has a direct connection to it
The switch then buffers packets which it can then forward to other nodes within the system - 52What does a Switch forwarding table store?MAC addresses of host (sender)
Interface packet was received into the switch
TTL - 53Describe the Switch table forwarding processWhen a frame is received, the switch compares the sender's (host) MAC address to the MAC address table, if no entry exists it is added to the table
The switch then compares the receiver (dest) MAC address with the table. If there is an entry, the switch forwards the frame out the associated interface. If there is no entry, the switch floods all the interfaces it has within the table - 54What is the difference between a router and a switch?Switch - link layer device (examine link layer headers), learn forwarding via flooding deals with MAC addresses
Router - network layer device (examine network layer headers), computes tables using routing algorithms, uses IP addresses - 55How do port-based VLAN's work?Switch ports are grouped so that a single physical switch operates as multiple virtual switches
Forwarding is done via routing - 56What are the two characteristics of port-based VLAN's?Traffic isolation - frames sent from ports X-Y can only read X-Y
Dynamic membership - ports have the ability to be dynamically assigned
Trunk port - carries frames between VLANs defined over multiple physical switches - 57Describe IPInternet protocol is a best effort protocol meaning packets can be lost, out of order, duplicated and delayed
It is also a connectionless network protocol - 58Describe fragmentationPacket splits into smaller fragments each with it's own IP header when original is too big for the underlying network
- 59What are the two types of fragmentation?Transparent fragmentation
Non-transparent fragmentation - 60Describe Transparent fragmentationA packet is split in to smaller fragments by a router
These fragments are sent to the next router which reassembles the fragments and combines them to form original packet
The next network does not come to know whether any fragmentation has taken place - 61Describe Non-transparent fragmentationA packet is split in to smaller fragments by a router
These fragments are not reassembled until the fragments reach their destination
Fragments can be routed independently on each other
Overhead of extra IP packet headers - 62What happens if a fragment is lost?The whole packet has to be re-transmitted
- 63How does checksum work?Calculated at the receiver
original sum + 1's complement of the sum
if != 0xffff then we have an error - 64What is the loopback interface?The loopback interface allows your computer to communicate with itself
Mainly used for diagnostics/maintenance
(IPv4 Address is 127.0.0.1) - 65What are the three types of special addressing used in IP and describe them?Unicast - packets delivered to a specific host
Broadcast - packets delivered to every node on the network
Multicast - packets delivered to a 'defined group' of hosts - 66How do we calculate the Round Trip Time (RTT)?Time of day - time from the data field (When the packet arrived)
- 67How does trace-route work?IP packet is sent with a TTL of 0, causing an error and the destination router to reply back with an error message, thus giving us their address
If there are multiple hops, TTL is increased by 1 for the amount of hops until we obtain the destination address - 68Describe the Maximum Transfer Unit (MTU)The maximum transmission unit (MTU) is the size of the largest network layer protocol data unit that can be communicated in a single network transaction
- 69What are the two methods of finding the MTU?1. Repeat the process used in traceroute, but reduce the packet size each time until we receive a 'port unreachable' error
2. Select packet sizes based on likely network types, compensating for size of IP header - 70What protocol maps names to IP addresses?Domain Name Service (DNS)
- 71What protocol maps IP addresses to MAC addresses?Address Resolution Protocol (ARP)
- 72What protocol maps MAC addresses to IP addresses?Reverse ARP (RARP)
- 73How does hop-by-hop work?
- 74Whats the use of an ARP table and what does it store?Allows hosts to recently used mappings
Stores;
IP addresses | MAC address
mappings along with a TTL - 75Describe 'Gratuitous ARP'?Updates are made to the ARP table of systems before they request it in order to update outdated information
Leads to address dupication - 76What does MPLS stand for and what is it's goal?Multi-protocol Label Switching
To provide high-speed IP forwarding using fixed length labels instead of addresses - 77Whats the difference between a standard router and a MPLS capable router?As MPLS routers forward packets solely based on the label value, this enables flexibility as they can use source AND dest (as oppose to IP, which only uses dest) to route flows to the same dest differently, this is called traffic engineering
- 78What does DHCP allow?A computer to join an IP network without having a pre-configured address
- 79What are the two DHCP components?DHCP Server
DHCP Client - 80What does each component do, respectively?DHCP Server is assigned to a specific network along with some configuration parameters and a database for manual IP acquisition and pools of addresses
DHCP Clients automatically retrieving DHCP settings - 81What is the DHCP procedure?DHCP Client broadcasts DHCP DISCOVER
DHCP Server then leases addresses to clients
After 1/2 the lease time, the DHCP Client can request to renew the lease, as long as the DHCP Client didn't time out they can get an extension of the same address - 82What are the three type of DHCP address leases?Manual lease - when network manager assigns all IP addresses
Automatic lease - when the DHCP Server permanently assigns some addresses, and dynamically assigns others
Dynamic lease - DHCP Server dynamically assigns IP addresses for a specific period of time - 83What are the pro's and cons of using DHCP?Pros;
Relieves network admin of manual config
Device can be move from N-to-N and automatically obtain value config params
IP addresses only allocated when needed
Cons;
When DHCP Server unavailable, Client unable to access network
Unreliable protocol
DNS can't be used - 84How does Network Address Translation (NAT) work?All datagrams leaving local network have the same IP address
All datagrams within local network have a static address, with nodes connected having different subnets - 85What is the function of the Transport Layer?Provide data transfer between processes at different end-points
- 86What are the two types of Transport layer service?Connection-oriented - consisting of set-up, data transfer and disconnect
Connection-less - the transfer of isolated units - 87What is TSAP an abbreviation for?Transport Service Access Point
- 88What are the functions of the Transport layer?Multiplexing/de-multiplexing data for multiple applications
Connection establishment
Error control
End-to-end flow control
Congestion control - 89What is a 'Port' in terms of address abstraction and how does it work?It is the concept of an abstract communication endpoint
It is identified by a positive integer
To communicate with a “foreign” port
Sender needs to know
IP address of the destination device
Protocol port number of the destination within that device - 90What do we use Port over PID for address abstraction?Processes are generated & terminated dynamically
Service to process isn't a 1:1 relationship - 91What is 'UDP'User Datagram Protocol
A simple transport protocol
-Best effort delivery over IP
-Unreliable
-Connectionless
-Message oriented - 92What are characteristics of UDP?No connection establishment
No QoS (quality of service) guarantees
No flow control
No error control/retransmission - 93What are the two distinct parts to internet 'routing'?Routing - application level process to determine routes
Forwarding - system level process directing packets according to learn routes - 94What is the process of IP forwarding and which layer does it occur?Occurs at the network layer
involves selecting the best known path to destination - 95How do we determine what class an IP address is?Using the first four bits of the host bytes
if first bit is 0 - Class A
if first 2 bits are 10 - Class B
if first 3 bits are 110 - Class C
if first 4 bits are 1110 - Class D
if first 4 bits are 1111 - Class E - 96How is the network/host divided among Class A, B and C?A - First byte network, last 3 bytes Host
B - 2 bytes network, 2 bytes host
C - 3 bytes network, 1 byte host - 97How does internal sub-netting work?An internal network is able to borrow host bits to form an internal sub-network
- 98What is VLSM and how does it work?Variable Length Subnet Masking
Subnet mask used to split the sub-net and host parts
Mask MUST have a set of ones followed by a set of zeros
Any IP address bitwise AND'd with it's subnet mask gives network address on which is resides - 99What is super-netting?The grouping of multiple contiguous address ranges of the same class
This is also called Classless Inter-domain routing (CIDR)
-Simplifies allocation of addresses
-Allows more flexibility - 100What are the two main aims of CIDR and what is the notation?Reduce number of routing table entries
Slow down the IPv4 depletion (RIP)
Address/prefix-length - 101Where are addresses bound and what are the basic parameters?Interfaces (i.e. routers)
Network address
Subnet mask
Subnet broadcast address - 102What is stored in a routers forwarding table?Network address/prefix
Subnet mask
Next hop router
Interface
Metric (cost) - 103Determining next hop (not a question, have an image to help me)
- 104What should you do if there are multiple next hop matches?Select the one with the longest prefix
this is because it should be the best path to the destination - 105What is route summarisation?The combination of two or more routes that only differ in the last bit
Basically we can replace .2 or whatever with 0/number of bits we have for the network allowing us to send to anyone on the sub-net - 106What are the risk when using a summary route?Holes, and these are routes which are no longer covered by the summarisation of the routes
- 107How can reduce routing table sizes?CIDR
Use of default routes - routes to direct traffic if all others fail - 108What are the two forms of IPv4 broadcast addressesLink-layer broadcast address - 255.255.255.255
Sub-net broadcast address - X.X.X.255
(Where X.X.X is internal sub-network) - 109What are the two types of IP delivery?Direct delivery - Use ARP to obtain MAC address of destination, send over connected link
Remote delivery - Find next hop in table using longest-prefix match, when at router before dest use direct delivery - 110What is RIP?Route Information Protocol
- 111How does routing/forwarding work with RIP?Routing processes exchange route updates via a routing protocol
- 112What are the two approaches to configuring routes using RIP?Non-adaptive/static - Relies on administrator establishing consistent routes on devices across networks under their control
Adaptive/dynamic - Routers run distributed algorithm, aim to establish best path between any two points - 113What are the two routing protocols use in RIP?Interior (LANs) -
Distance vector
Link state
Exterior (WANs) -
Path vector - 114Explain what the Distance Vector interior routing algorithm doesDistance vector requires that a router inform its neighbors of topology changes periodically (what nodes it is able to get to)
Based on calculating the direction and distance to any link in a network
Routers exchange {destination, metrics} pairs (metrics = cost, distance etc) - 115What is a problem caused by the Bell-ford algorithm?Count-to-infinity
This problem is cause when, lets imagine we have a small network of nodes (A,B,C,D)
A has a route to C via B, but B has gone offline A still things that B is a viable route and tries to send data that way, inevitably causing a routing loop (congestion), making C & D unreachable (in this case) - 116How is this problem combatted?If an increased cost is received via a link we've already got recorded, we must update the cost as we can assume link failure, thus allowing other paths/links to be an option resulting in the system being fully functional again
- 117What is the 'Bouncing Effect'?Each full round the network adds double the link cost to related entries
This causes packets to loop around the network, causing congestion, this ultimately resulting in packet loss - 118How often are RIP updates sent, and what is the recommended TTL?Every ~30 seconds
6x ^ - 119What are 'Triggered Updates'?Nodes send updates early if there is a change in their route table
Either due to link failure/increased link cost - 120What is 'Split Horizon'?An observation which says;
If X has route to Z via Y, Y must never try go to X via Z
Requires different updates on each adjacent link - 121What's different about 'Split Horizon with Poisoned Reverse' as oppose to the original?Instead of not allowing packets to return, just return them with a cost of infinity (i.e. explicitly say that it's unreachable)
- 122What is a 'Solicited Update'?Device sends RIP-REQUEST
Peer replies with RIP-RESPONSE (can be multiple), giving a list of known routes - 123What is an 'Unsolicited Update'?Regular RIP-RESPONSE's being broadcast, they're also sent on significant connectivity changes (i.e. link failure)
Can be a problem if all routes do this at the exact same time - 124What does RIPv2 support that RIPv1 didn't?CIDR
Multicast based updates - 125What IGMP and what does it do?Internet Group Management Protocol
Used between device and first-hop router and between devices on the same link - 126What are problems with RIP?Simple - no support for dynamic metrics
Robust - info good/bad freely distributed
Scalability - doesn't scale for the whole internet
Slow convergence - updates are sent node-by-node
Loops - 127When is an 'Acknowledgement' returned?When the receiver gets the correct packet
- 128What does ARQ mean and what do these protocols all have?Automatic Repeat ReQuest
-Error detection (Checksums)
-Receiver feedback (ACK, NAK)
-Retransmission of corrupt packets - 129How does the 'Stop and Wait' protocol work?Sender waits for ACK before sending another packet
IF corrupt and don't know whether ACK/NAK
Assume ACK - send new
if ACK - OK
NAK - missing packet
Assume NAK - resend old
if NAK - OK
if ACK - duplicate packet - 130What are sequence numbers?one bit sequence of numbers that allows both ends to spot peer 'misunderstanding'
- 131What is the 'Alternating Bit Protocol'Sequence numbers that switch between 0 and 1 (ACK0, ACK1)
- 132How do we select a timeout value?If it's too short, we'll resend unnecessarily causing duplicate packets
If it's too long and any problem that occurs will give long delay
Value needs to be at least RTT - 133How do we calculate the transmission time of a packet?packet length / transmission rate
- 134How do we calculate utilization time?Total time = Dtrans + RTT
Utilization time = dtrans / total time - 135Why do we 'Pipeline'?Allows increased range of sequence numbers
Allows the sending of packets until no longer have sequence numbers (a lot more efficient than stop and wait) - 136What defines the range of sequence numbers required?Transmission speed - Faster -> more sequence numbers required
RTT - Longer -> more sequence numbers required - 137What is the idea behind the 'Slide Window' protocol?That we have a range of sequence numbers that;
Correspond to a packet which is being sent and has been delivered ok
Correspond to packets which have been sent, but not ACK'd
Correspond to available to be ACK
Correspond to some unavailable - 138What is the 'Window Size'?Number of sequence numbers that are able to pipelined
When ACK's received the window slides along - 139What is the window size limited by?Desire to limit un-ack'd packets
Number of bits for sequence numbers - 140What is 'Go-Back-N'?Is an ARQ protocol that returns to the sequence number of the corrupt packet once it times out, then resends from that point
The greater the seq numbers the more this becomes an issue - 141What are 'Cumulative Acknowledgements'?ACK's for all packets up to a point, e.g. ACK7 would mean all packets up to this point would have been ACK'd
Eliminates the needs to ACK each packet - 142What is 'Selective Repeat'?Receiver - Individually ACK's all packets, slides forward only when no missing messages
Sender - Slides window forward once all ACK's to that point received - 143What are the reliability mechanisms?Checksums
Timers
Sequence numbers
Acknowledgements
Sliding windows
Pipelining - 144What is DSLDigital Subscriber Line
Uses telephone line to exchange DSLAM (Digital subscriber line access multiplexer)
Data - ISP/internet
Voice - telephone net - 145What is ADSL?Splits bandwidth on the wire into uplink and downlink parts
e.g. 256kb/s up, 2mb/s down
Permanently connected - 146What is DCN abbreviated for and what is the structure?Data Center Networking
Load balancer - application-layer routing
Receives external client requests
Directs workload within data center
Returns results to external client - 147What does FTTC and FTTP meanFibre to the Cabinet
Fibre to the Premises - 148Why do we need IPv6?32 Bit addresses too small
0 unallocated IPv4 addresses left
Routing table size has become a big problem
Management is overfly complex - 149What are the two main cases for IPv6?Security - no network layer support, difficult obscure end-points and extensions are often unsupported
Mobility - extensions are overly complex, rarely support, ineffcient - 150How are IPv6 addresses represented?128 bits long (CIDR style prefix)
written in hex
xxxx: x8/2
Leading zeros can be omitted - 151What are some notable IPv6 addresses?0.0.0.0.0.0.0.0 - Unknown address
0.0.0.0.0.0.0.1 - Loopback address - 152How do we map a 48 bit MAC address to EUI-64 format?Take the MAC address, and insert FFFE in the middle, then take the first byte (from the left)and convert it to bits, then invert the 7th bit from the left
- 153How does Stateless Address Auto-Configuration work?1. Map MAC -> EUI-64
2. Form link local address FE80::
3. Sends router solicitation message
4. Waits for router advertisement - 154What are the two types of IPv6 private addresses?Link-local
1111111010 - 54 bits for subnet - EUI-64
Site-local
1111111011 - 54 bits for subnet - EUI-64 - 155What addresses do all IPv6 nodes have?Link-local
Unicast/Anycast (not mandatory)
Loopback - 156What is DAD?Duplicate Address Duplication
Duplicate addresses can seriously disrupt communication
Must be performed on all normal addresses
-Link-local addresses
-Global unicast addresses - 157How do we test a new network protocol?Tunneling/packet encapsulation
- 158How does 'Tunneling' work?Pair of routers maintain an IP 'Tunnel'
Add/Remove extra IP header for packets to be routed over link
Works by IPv6 packet being encapsulated by a IPv4 header - 159What are FQDN's?Full Qualified Domain Names
end with a period
e.g. cs-lab.co.uk. - 160What is the 'Directory Service'?Service for name/address look up
- 161What are the issues with the 'Directory Service'?Scalability - has to process large amount of queries
Availability - fault isolation & recovery - 162What is the 'Domain Name Service' (DNS)?A distributed database
All domains should have one primary server, the reliable name server and one secondary server somewhere else in the world - 163Why is the database distributed?No central point of failure - secondary servers mirror primary servers
Reduce load on the root servers
Local servers - reduce request latency - 164What are the characteristics of DNS?Scalability - no limit to the size of the DB, number of queries
Reliability - data is replicated (copied from primary), clients can query the primary and secondary servers
Dynamicity - DB can be updated dynamically, modification of the master DB triggered replication i.e. backup - 165How do domain names work?Sequence of labels from a node to the root by its subdomains
One domain is a subdomain of another if domain ends in others domain name e.g. ac.uk subdomain of .uk - 166How does 'Delegation' work?Admins can create subdomains to group hosts
An administrator of a domain can delegate responsibility for management of a subdomain to someone
The parent domain retains links to the delegated subdomains - 167How do 'Zones' work?Each time an admin delegates a subdomain, a new unit of administration is created
Allowing subdomain and parent domain to be administered independently - 168What is the process of 'Name Resolution'?It's the process by which resolvers and name servers cooperate to find data in the name space
- 169What are the two things we need to resolve an address?Local host file
DNS - 170What is a forward/reverse look up?Forward loop up
Domain name -> IP address
Reverse look up
IP Address -> Domain name - 171What is a 'Glue record'?The association of a hostname with an IP address
- 172Whats the use of secondary servers?Provide 'backup' service and reduce the load
Few secondary servers for each domain - 173What is the DNS process?Resource records generated by zone's primary server
Secondary servers hold copies to protect against failure
Clients contact their local server via resolver
Cached records can be returned as replies to upcoming queries - 174What are some of the attacks DNS faces?DDoS
Hijacking
Spoofing - 175How is a DNS DDoS attack carried out?DNS servers are flooded with requests until they fail
- 176How is a DNS Hijack carried out?Operating system/web browser becomes infected with a virus/trojan
Intercept communications and direct you to a false site - 177How is DNS Response spoofing achieved?Eavesdrop on requests, outrace the servers response
- 178What is TCP?Transmission Control Protocol
Stream based transport protocol - 179What are the key features of TCP?Stream oriented - bytes delivered in order
Virtual circuit based
Buffered transfer - transfers bytes as network capacity allows
Unstructured - transfer independent of application data structures
Reliable
Full-duplex - 180Describe the 'Three-way Handshake'1. A->B SYN - My Seq Num is X
2. B -> A SYN ACK - Yours X, Mine Y
3. A->B ACK - Your Seq Num is Y - 181Whats the correct way to close a connection?Client sends FIN flag
Server receives flag, responds with ACK
Server sends FIN flag
Client receives flag, responds with ACK then enters TIME_WAIT state until server closes connection - 182How do you calculate the 'Maximum Segment Size'?MSS = MTU - headers
- 183How are retransmissions calculated with TCP?EstimatedRTT = a x EstimatedRTT + (1-a) x SampledRTT
Timeout = B x EstimatedRTT
a = 0.8-0.9
b = 2 - 184How does TCP handle congestion?TCP assumes packet loss = congestion
TCP has a 'congestion window'
Aggressively reduces traffic if detects congestion - 185How does 'Multiplicative Decrease' work?For every lost segment (timeout) we halve the congestion window to a minimum of the MSS and back-off retransmission timer exponentially
If continued loss, the congestion window will reduce exponentially - 186How does 'Additive increase' work?After congestion;
Start congestion window at 1 segment (the MSS)
Increase congestion window by
1 segment with each congestion window's worth of ACK'd segments - 187What is the 'Slow Start' congestion strategy?Send data at the receivers advertised window
There is zero consideration for intermediate devices, causing congestion - 188What does the 'Fast Retransmit' enhancement reduce?The amount of time the send waits before retransmitting a lost segment
- 189What are the two types of TCP?TCP Tahoe
TCP Reno - 190How does 'Fast Recovery' work?When TCP performs a fast retransmission each duplicate ACK must indicate a segment leaving the network, enables us to maintain level by sending one segment for each ACK
- 191How does TCP Tahoe work?TCP Tahoe - when a loss occur, fast retransmit is sent, half of the current congestion window is saved as a Slow Start threshold. Slow start now beings from the initial congestion window until it reaches the Slow Start threshold then uses a 'congestion avoidance algorithm'
- 192How does TCP Reno work?TCP Reno - implements the 'fast recovery' algorithm. A fast retransmit is sent, half of current congestion window is saved as the slow start threshold AND as the new congestion window, thus skipping slow start and going directly to Congestion Avoidance algorithm
- 193What is OSPF?Open Shortest Path First
- 194What are some interior gateway protocols?RIP
IS-IS
OSPF - 195What are some advantages of OSPF?Connects quickly without loops
Supports multiple precise metrics simultaneously
Can load-balance over multiple paths to destination - 196How does the link state routing algorithm work?Discover neighbours and their addresses
find metrics to each neighbour
build announcement packet
send to all routers
compute shortest path algorithm - 197What's the difference between link state and distance vector routing algorithms?Link state - announce to all nodes, learns link cost of immediate neighbours. Each node learns full network structure. All nodes compute all shortest paths independently
Distance vector - announce to immediate neighbours, least cost distance to all nodes, distributed computation of shortest paths - 198How does OSPF work on Broadcast networks?Every router would share updates with every other
Election of a designated router - meaning all routers synchronise with this device - 199What does a backbone provide in OSPF?The avoidance of distance vector problems
- 200What is stub area?Area only having default route for external networks
Informationen
- Beschreibung
- Karteikarten am SCC 203 - Networking, erstellt von Mac Howarth am 04/05/2017.
- Erstellt von
- Mac Howarth
- Aufrufe
- 4
- Bewertung