MOD 5 - Network Layer - Control Plane SDN

Question 1

In the Network layer, what are the 2 main functions?

Answer 1

1) DATA PLANE = Forwarding : move packets from router’s input to the appropriate router output
2) CONTROL PLANE = Routing : determine route taken by packets from source to destination

*note: every router has a control and data plane!

Question 2

What are the 2 ways to structure the network control plane (route taken by packets from src to dst)

Answer 2

1) per-router control (traditional) = Individual routing algorithm components in each and every router interact with each other in control plane to compute forwarding tables
2) logically centralized control (software defined networking SDN) = A distinct (typically remote) CONTROLLER interacts with local control agents (CAs) in routers to compute forwarding tables.
In this approach, a central entity makes decisions and controls the behavior of the entire network.

Benefit: The controller can develop a consistent view of the network state for calculating shortest paths and can implement application aware routing.

*see pics in notes

Question 3

what is the goal of routing protocols

Answer 3

To determine “good” paths (routes), from sending hosts to receiving hosts through a network of routers.

*Path = sequence of routers that packets will traverse through from given initial source to given final destination host.
*“Good” conditions = 1)lowest cost, 2) fastest, 3) least congested paths

Question 4

What is the purpose of routing algorithms?

Answer 4

To find the least costly path between routers.

Question 5

Link state algorithms vs Distance vector algorithms?

Answer 5

Link state algos = CENTRALIZED, global info, All routers have complete topology & link cost info. Represented by vector of link cost to neighbors ex: A{B:2, C:3} *A links to B and C
“tell the world about the neighbors”

Distance vector algos = DECENTRALIZED, only neighbor info, Router knows physically- connected neighbors & link costs to neighbors.
Iterative process of computation, exchange of info with neighbors.
Represented by vector of distances to all nodes ex: A{A:0, B:2, C:3, D:4, E:10} *all distances from A
“tell the neighbors about the world”

Question 6

What is flooding and what’s its purpose? When do we initiate it? What are some challenges and their solutions?

Answer 6

Flooding = router sends link-state information out its neighboring links, then next router does same.. Until the last router.

Purpose = to ensure all router have the latest link-state info

Challenges & solutions = packet-loss & OoO arrival -> fix with ACKs and retransmission, Seq numbers, and TTL for each packet (remaining #hops a packet do)

Initiate when : there’s a topology change (link or node change/failure), there is a configuration change (link cost change), periodically (to refresh, link-state and prevent corrupt data)

Question 7

Distance vector algorithm uses which equation? What’s the the equation?

Answer 7

Bellman-Ford equation
dx(y) = cost of least-cost path from x to y
= minv {c(x,v) + dv(y)}

x = origin, y= dest, v=all neighbours/links of origin x
Minv = taken over all neighbors v of x
c(x,v) = cost to neighbor v / link x-v cost
dv(y) = cost from neighbor v to destination y

Question 8

In DV algorithms, what is the count to infinity problem?

Answer 8

Scenario: The problem arises when there is a change in the network topology (ex: link failure). In DV algorithms, when a router detects a change, it updates its routing table and informs its neighbors of the change. Then the neighbors update their tables and sent the information to their neighbors…etc.

Issue with this: It takes time for these updates to propagate through the network. During this propagation time, routers may make decisions based on outdated information, potentially leading to incorrect routing decisions (ex: infinite loop)

Question 9

Compare Link state and Distance vector algos.

Answer 9

Message complexity
LS =exchange between all nodes (a lot) O(n^2)
DV = exchange between neighbors only O(n of neighbours)

Convergence
LS = O(n^2), slow
DV = convergence time varies
Possible routing loops
Count-to-infinity problem

Robustness
LS = Better!
Node can advertise incorrect LINK cost
Each node computes only its own table
DV =
Node can advertise incorrect PATH cost
Each nodes table is passed on to other nodes = errors can propagate through the network

Question 10

What is a AS region (Autonomous system)?

Answer 10

Region of grouped/aggregated routers

Question 11

What is intra-AS routing?

Answer 11

Routing among hosts/routers in the SAME AS network/domain
Rules within the AS: all routers within the AS must run the same intra-domain routing protocol

*see drawing if needed

Question 12

What is Inter-AS routing?

Answer 12

Routing amongst different AS’s network/domain

UNIQUE protocol = BGP

*see drawing if needed

Question 13

What are gateway routers in AS’s?

Answer 13

Router at the edge of a AS that has a link to another router in another AS.
They also perform inter-domain routing (routing from AS to AS) and intra domain routing (routing only between the same AS)

*see drawing if needed

Question 14

Name common Intra-AS routing/ interior gateway protocols (IGP)

Answer 14

RIP, Routing Information Protocol (classic DV)

OSPF, Open Shortest Path First (classic link-state routing)

IS-IS (identical to OSPF)
IGRP
EIGRP

Question 15

In intra-AS routing, OSPF is uses ___ routing algo and RIP uses ____routing algo. What is the goal of of these intradomain routing algos?

Answer 15

OSPF = link state routing (dijkstra)

RIP = DV, distance vector routing

The goal of an intradomain routing algorithm such as link-state or distance-vector routing is to enable routers within the same autonomous system to efficiently exchange information about network topology and calculate optimal paths for forwarding traffic in that system.

OSPF allows for summarising distances to nets in its own area
BGP allows subnets to advertise their existence to the rest of the Internet.

Question 16

What is BGP, iBGP and eBGP (interdomain)?

Answer 16

= Border Gateway Protocol, provides each AS a means to
1) Allows subnet to advertise its existence to rest of internet
2) Determine “good” routes to other networks based on reachability information and policy

2 types:
1) External BGP (eBGP) = obtain subnet reachability information from neighboring AS’s. Gateway router to gateway router connection.
2) Internal BGP (iBGP) = propagate reachability information to all AS-internal routers. Router to router connection in same AS (to det best GWR connection).

Question 17

what is hot potato routing?

Answer 17

to calculate NEXT-HOP: chose local gateway that has least intra-domain cost (link cost)

Question 18

What are the important factors to look at in BGP routing?

Answer 18

1- The reachable network prefix: policy decision

2- AS-PATH = list of AS’s through which prefix advertisement has passed

3- NEXT-HOP = : IP address that indicates specific internal-AS router to next-hop = AS, hot potato routing

Question 19

Intra-AS routing vs Inter-AS routing, in terms of policy, scale and performance?

Answer 19

Policy:
Intra-AS = single admin, so no policy decisions needed
Inter-AS = admin wants control over how its traffic routed, who routes through its net.

Scale:
intra-AS: Since it’s a single administrative domain, the routing system doesn’t need to handle the complexity of multiple domains.
inter-AS: hierarchical routing is needed as it saves table size, reduced update traffic

Performance:
intra-AS: can focus on performance (as there’s no policy)
inter-AS: policy may dominate over performance

Question 20

What is ICMP? Name 2 different ICMP applications

Answer 20

= internet control message protocol

ICMP messages are carried in IP datagrams. Operates at network-layer.

Used by hosts & routers to communicate network-level information
This includes:

• error reporting: unreachable host, network, port, protocol
• echo request/reply (used by ping) (used to test reachability and rtt of the network specific to a dst host)

applications:
-ping
-tracerouter

Question 21

What is network management?

Answer 21

Summarized as FCAPS
F= Fault management
C= Configuration
A= Accountability
P= Performance management
S= Security

Question 22

Compare how well link-state algorithms and distance vector algorithms respond in the event of a router failure

Answer 22

Link-state algorithms typically respond more efficiently to router malfunctions, in comparison to distance vector algorithms.

With link-state algorithms, if a node advertises an incorrect link cost, errors are localized, and each node computes only its own routing table.
On the other hand, with distance vector algorithms, if a node advertises an incorrect path cost, errors can propagate through the network since each node’s routing table is used by others.

Question 23

Suppose a network uses distance vector routing. What would happen if a router sent a distance vector with all zeros?

Answer 23

*not sure
If a router using distance vector routing sends a distance vector with all zeros, it communicates that it has a route with no cost to reach the destination, and that the associated route is unreachable.

Question 24

Describe how loops in paths can be detected in BGP.

Answer 24

When a BGP router receives an update, it will check the AS-PATH to ensure that its own AS number is not present in the path. If the router’s AS number is already in the AS-PATH, it will indicate a loop. In result, the router will reject the update.

BGP has built‐in loop detection via the AS_PATH attribute. The loop can be detected if an AS appears more than one times

Question 25

In OSPF, what is a border area router, backbone router, boundary router?

Answer 25

area border routers: “summarize” distances to nets in own area, advertise to other Area Border routers.

backbone routers: run OSPF routing limited to backbone.

boundary routers: connect to other AS’es.

Question 26

what are stub networks?

Answer 26

network segment that has only one route in or out. It typically has one entrance/exit point with not alternate option.

Question 27

What is SNMP?

Answer 27

= Simple Network Management Protocol
- is an application layer protocol mostly run on top of UDP. Used to manage and monitor network devices.
- It provides a standardized framework and a set of rules for managing network devices.
- SNMP organizes management information in a hierarchical structure called the Management Information Base (MIB)

Question 28

T/F? BGP exchanges link weights.

Answer 28

False; BGP exchanges path vectors

Question 29

how does BGP solve the count to infinity problem?

Answer 29

BGP is a path vector protocol, and it operates based on the information of the entire path to a destination (called announcements), rather than just the distance or cost.

A router, upon receiving a BGP announcement, can check for its own AS number in the path; if the AS number is present, the path has a loop and is immediately discarded.

Question 30

What scales better, D-V or L-S routing?

Answer 30

D-V

Routers in distance vector protocols only receive updates from their immediate neighbors, while routers in L-S must manage updates from every router in the network.

This makes DV routing easier to scale

Question 31

When routers generate ICMP messages, to where do they send them? Along with the ICMP
header at the beginning, what additional contextual information do routers include in the
messages?

Answer 31

ICMP messages are sent to the source address specified in the IP header.

ICMP messages contain
- ICMP header
- IP header of the packet that triggered the ICMP
- along with at least 8 bytes of that packet’s payload, which is enough to include the ports used in the transport header if the packet was carrying UDP or TCP.

(ICMP header[ type field, code field], plus first 8 bytes of IP datagram causing error)

Question 32

Are ICMP messages reliable?

Answer 32

No, since they are delivered in IP datagrams (network layer).

This means that they are treated like any other IP packet, thus also subject to packet loss.

Question 33

What do the following ICMP messages do/mean:
- ping (echo request/reply)
- dest host unreachable
- dest port unreachable

Answer 33

• echo request/reply = testing reachability of a specific host, request = to send ping, reply = answer after you receive ping
• dest host unreachable = can’t reach the network specified
• dest port unreachable = UDP port you are trying to reach has no socket associated with it

Question 34

How would an SDN enabled router differ from a traditional router?

Answer 34

• Network switches become simple forwarding devices
• SDN separates the control plane and the data plane
• SDN has a logically centralized controller

Question 35

In LS routing protocols, what is the main problem?

Answer 35

• each node computes only its own routing, overhead with exchanging & maintaining link state info

Question 36

How would you perform shortest path routing in a very large network?

Answer 36

• Hierarchical OSPF, divides network into smaller areas. Each area has its own OSPF domain, and routers in that area have detailed network knowledge

Question 37

How can you use ICMP messages to detect a lost link along a routing path as a network administer?

Answer 37

by using ping or traceroute commands

Question 38

Briefly describe how the traceroute tool works (i.e., what does it do in order to identify the routers that make up an Internet path)

Answer 38

• sends ping requests to identify the routers in a path from src to dst.
• sends out messages (UDP segments) with a countdown (TTL) to find the destination. When a message reaches a router on the way, that router sends a signal back (ICMP message) saying, “you’re getting warmer,” and the source notes the time it took (TTL=1,2,3..). This keeps happening until the final destination says, “found you,” and the game stops. It helps us see the path and time it takes for data to travel across the internet.

Question 39

In hirerachical OSPF, what is a 2-level hirerachy, area border routers, backbone routers, boundary routers?

Answer 39

▪ two-level hierarchy: local area, backbone.
* link-state advertisements only in area
* each nodes has detailed area topology; only know
direction (shortest path) to nets in other areas.

▪ area border routers: “summarize” distances to nets in
own area, advertise to other Area Border routers.
▪ backbone routers: run OSPF routing limited to
backbone.
▪ boundary routers: connect to other AS’
es.

Question 40

Discuss how a hierarchical organization of the Internet has made it possible
to scale to millions of users.

Answer 40

Routers are organized into autonomous systems (ASs). Within an AS, all routers run the same intra-AS routing protocol. The problem of scale is solved since a router in an AS need only know about routers within its AS and the subnets that attach to the AS. To route across ASes, the inter-AS protocol is based on the AS graph and does not take individual routers into account.

Question 41

Compare and contrast the advertisements used by RIP and OSPF.

Answer 41

OSPF= linkstate: periodically broadcasts routing information to all other routers in the AS, not just to its neighboring routers
RIP = distance vector: contains information about all the networks in the AS, although this information is only sent to its neighboring routers