Questions tagged [congestion]
For questions about network congestion, where a network node or link is carrying more data than it can handle. For instance, you are troubleshooting or trying to find a bottleneck in your network topology.
75
questions
0
votes
1
answer
34
views
What does "Configure kernels to not follow the RFC 5681 SHOULD" mean?
Netflix documentary The Playlist about Spotify, they present one of the technical breakthroughs that they created their own TCP implementation to reduce the latency from first click to playing the ...
- 101
0
votes
0
answers
34
views
Congestion Window Equation
I'm reading through High Performance Browser Networking and would like to understand how this equation for congestion window sizing works. I'm not sure if this would be more appropriate on a math ...
- 1
0
votes
1
answer
58
views
Does TCP CUBIC reduce window size for an idle connection?
Does the TCP CUBIC congestion control algorithm reduce window size for a connection that has been idle for a time but has not yet timed out? For example, if an application starts sending data and ...
0
votes
1
answer
146
views
What is the unit and the behaviour of the congestion window?
I've learned that the congestion window is the maximum amount of packets that can be sent in one transmission round.
Now TCP seems to acknowledge the packets in bytes, so wouldn't it make more sense ...
- 1
0
votes
2
answers
66
views
Why is congestion collapse from undelivered packets not possible when there's 1 congested link?
rfc 2914 says this:
The avoidance of congestion collapse from undelivered packets
requires that flows avoid a scenario of a high sending rate, multiple
congested links, and a persistent high packet ...
- 31
1
vote
1
answer
210
views
What stategy use TCP Cubic in Fast Recovery?
I want to do some simulations with different TCP algorithms I was reading one of the earliest papers for TCP CUBIC and I don't see explicitly the strategy of Fast Recovery after a loss of the packet.
...
1
vote
2
answers
229
views
Differences between TCP congestion and TCP flow control
From my understanding, flow control essentially how much the receiver can process.
Suppose I can send packets at 100mb/s but the receiver can only process information at 10mb/s, flow control would ...
- 13
0
votes
1
answer
81
views
"Expected" packet loss under heavy network load in TCP cubic
context
We have two machines that receive a lot of data, and mirror each other's content with rsync. When monitoring these boxes via ICMP echo requests (using the mtr tool), we see up to a couple % of ...
- 3
1
vote
1
answer
159
views
Why does TCP sender drop rate 50% on lost-packets?
Why is the standard TCP sender response to packet-loss (congestion), to drop 50% of the sending rate and then start working back up ?
If packet-loss was 20%, why not just drop 20% or maybe 25% to be ...
- 111
0
votes
1
answer
80
views
In what situation would normal choke packet be preferred over hop-by-hop choke packets?
When the host is far away, it makes sense to use hop-by-hop choke packets to resolve congestion.
But is there situations where normal choke packet is more appropriate?
And by the 'source' we mean a ...
- 1
3
votes
2
answers
2k
views
How is congestion avoided when using UDP?
I understand how TCP goes about this - through various methods such as Congestion Window (CWND), Sliding Window, Slow Start and Fast Recovery - it's basically built into the protocol. I understand ...
- 133
1
vote
1
answer
72
views
Why does congestion happen
Imagine a network full of routers . The job of a router is to be able to receive multiple packets from multiple sources then send the packets 1 at a time in each of its outputs.If the routers are able ...
- 113
0
votes
2
answers
430
views
Does window size in tcp header include tcp header size?
Let's say a TCP segments has in the field of "window size" = 1400 bytes.
Does that mean that
we can send 1400 bytes counted in tcp segments = TCP headers + data
or
it does specify only ...
- 103
-1
votes
1
answer
45
views
Latency between a churn and congestion event triggered in the TCP code
I am analyzing the time it takes between a churn occurring in the network (e.x. that causes packet drops or delays) and when the congestion event handling codes are triggered in the TCP kernel in the ...
- 35
-1
votes
1
answer
749
views
Why sender window or congestion window size becomes 2^n mss after nRTT in slow start mechanism?
I seen everywhere in slow start mechanism if initial sender size denoted by cwnd= 1mss , then after 1 RTT cwnd becomes = 2 mss, and after 2 RTT cwnd becomes = 4 mss and so on. My question is why ...
- 378
2
votes
2
answers
502
views
Congestion control behaviour of TCP Reno
Which behaviour of TCP Reno after 3 DUP ACK is correct:
The one described in the Italian Wikipedia (https://it.wikipedia.org/wiki/Controllo_della_congestione_in_TCP#TCP_Reno). It says the new cwnd is:...
- 21
0
votes
0
answers
37
views
Why large queueing delays are experienced as a packet arrival rate nears the link's capacity? [duplicate]
From the book "Computer Networks A top Down Approach":
It says if hosts A and B both are sending at rate R/2 and router's outgoing link has a capacity of R. There will be delays experienced ...
- 25
1
vote
1
answer
222
views
Does UDP's lack of congestion control clog up the Internet?
I'm reading Computer Networking: A Top-Down Approach, and from what I understand, UDP has no widely adopted mechanism for congestion control like TCP does. Isn't this a problem for the Internet as a ...
- 113
0
votes
1
answer
162
views
Difficulty in understanding the concept of "artificial inflate" and "deflate" in the fast recovery phase of TCP
I was going through the text Computer Networking- A Top-Down Approach by Kurose and Ross, there I found subtleties with the TCP congestion control FSM which is shown below:
Mainly I am having ...
1
vote
1
answer
691
views
Tearing down the FSM description of congestion control
I was going through the text Computer Networking- A Top-Down Approach by Kurose and Ross, there I found subtleties with the TCP congestion control FSM which is shown below:
Mainly I am having ...
1
vote
1
answer
45
views
tcp how the receiver behaves if it received an incomplete window
Suppose, for example, that the window has a size of 5 MSS, the sender sends 5 segments to the network, the fifth is lost. Actually the question: how does the recipient behave? It just doesn't ...
- 113
-1
votes
1
answer
32
views
Would a bidirectional exchange of data between two hosts, one using TCP Reno and the other using TCP Tahoe, be possible?
I know Reno and Tahoe are TCP congestion control algorithms and one recovers much faster than the other when a packet is lost, but would they affect how data are exchanged between the hosts? ...
- 3
0
votes
2
answers
236
views
Congestion control in transport and network layers in TCP/IP
Why do we need congestion control in BOTH the transport layer and network layer in TCP/IP model. Cant we be good enough with control in just one of the layers only? And how do they differ?
- 3
2
votes
1
answer
325
views
TCP Congestion Control, Sender window = minimum ( Congestion window, Receiver window)
When can the congestion window size be greater than the receiver window size? An example of a scenario will be helpful.
I know that the congestion window is used by the sender to gauge the network ...
- 23
0
votes
1
answer
23
views
Does higher retransmission time imply lesser chances of congestion?
I believe that with higher retransmission time value, the frequency at which retransmissions occur will reduce. And with less frequent retransmissions, the traffic will decrease. So with that logic, I ...
- 113
1
vote
1
answer
191
views
TCP flow control active waiting
To perform flow control between a sender and a receiver during a TCP connection, the receiver includes in each packet to the sender, the size of available space ‘rwnd’ in its receiving buffer. This is ...
- 11
0
votes
1
answer
125
views
Effect of block size on network congestion?
I am interested in comparing TCP/IP network congestion under two scenarios:
Scenario 1: A large number of senders send messages to a group of receivers
Scenario 2: Same as above except that the ...
- 103
0
votes
0
answers
77
views
Reason for congestion in CSMA based networks
A typical channel with SNR 10dB and BW 10 MHz will have a channel capacity of something around 150 Mbps. In CSMA, at any given time, only 1 user is allowed to transmit and in 802.11p, the default ...
- 101
0
votes
1
answer
3k
views
TCP cwnd is increasing when congestion occur
I am a newbie network engineer.
I did a small experiment with iperf3. But I run into a small problem.
(1) My question.
I think TCP flow do not act properly despite of its congestion control.
The flow ...
- 25
0
votes
1
answer
186
views
Response time of TCP congestion control
I would like to know about the overall latency of the TCP congestion control for adapting to varying network conditions. Specifically, my questions are:
How fast can the system react after a change ...
- 133
1
vote
1
answer
499
views
Routers and output buffers: Why can two devices interfere with each other and cause latency?
I apologize if this question has been asked already because I couldn't find anything related. I'm a beginner in networking and am reading through Computer Networking: A Top-Down Approach, as well as ...
1
vote
1
answer
200
views
What does the speed of Network Switch actually mean?
Network switches usually come in speeds of 10/100/1000 Mbps, and that's all fine and dandy, then I got thinking..
Suppose I have an 8 port 1000 Mbps switch, each port connected to a client. If a ...
1
vote
1
answer
406
views
Initial TCP congestion window sizes
I have tried to look for this information on Google, but I haven't been able to find a good source for this information.
So, I am interested in what are the initial TCP congestion windows on ...
- 125
1
vote
0
answers
367
views
Simple Token Bucket - Policing - Terms - CIR, CBS
I have gone through token bucket algorithm. However, I am struggling to apply this knowledge into the real devices like CISCO, etc. So, for QoS policing, interface meter. I am trying to understand ...
- 431
5
votes
2
answers
1k
views
QoS shaper, shapes traffic without dropping packets?
I have a problem with my QoS policy, or I should better say, I have a problem understanding how my policy manages to achieve what ask of it without dropping a packet!
Here is the scenario:
We have ...
- 53
4
votes
1
answer
2k
views
Relationship between TCP window size and TCP congestion control
Am I correct that TCP receive window(which is a send window for the sender) has nothing to do with TCP CUBIC congestion control? TCP receive window is solely a property of TCP flow control?
- 1,217
2
votes
2
answers
3k
views
What is congestion collapse?
My understanding is that it involves multiple copies of the same packet existing in the network. But how does that situation end up happening? And why is it so harmful?
- 127
2
votes
2
answers
511
views
AIMD: Total number of packets sent
While answering to the quizz 4-3 of the great Introduction to Computer Networking (Stanford Lagunita), I found myself unable to explain the answer of one of the questions:
Consider a long flow being ...
- 35
3
votes
1
answer
5k
views
Calculating average size of congestion window
In the computer networking book (James F. Kurose, Keith W. Ross), it is written that we need 83,333 in-flight segments(window size) in order to reach 10 Gbps throughput.
(We have 1500 byte segments &...
- 33
1
vote
2
answers
453
views
BGP: Route congestion
Using a BGP Multihoming design, is it possible redirect or modify any metric to determinate destination?
Example: I detected that the most of users are having problem to access a webserver: high ...
1
vote
1
answer
94
views
congestion control
Suppose , a sender sends a packet and it does not receive an acknowledgement . The timer at the sender side runs out . In which cases the sender resends the packets and in which cases it does not ?
...
- 27
1
vote
2
answers
3k
views
When should TCP send a cumulative acknowledgement?
I am simulating TCP and I don't know when it should send the Acknowledgment.
For example as soon as a segment dropped should we send the Acknowledgment or should we take some time, maybe the segment ...
- 11
1
vote
2
answers
1k
views
Ethernet Flow Control vs Quality of Service
What are the trade-offs/major differences of Ethernet Flow Control vs Quality of Service? How to know when to choose one over the other?
They seem to both be a solution for congestion.
The obvious ...
- 614
2
votes
1
answer
2k
views
Congestion control,flow control, MTU and MSS, Rwnd-Cwnd and slow-start
OK, so basically I'm getting confused with all the terms, and if some are related to others or not.
I'll explain myself:
From what I understand, flow control depends on the receiver side buffer ...
- 121
1
vote
1
answer
555
views
Can tcp throughput be equal to bandwidth?
I have a conceptual question which I cannot understand. Can the TCP throughput be equal to the bandwidth and why?
The throughput from what I have learned so far is (3/4)*congwing/RTT.
Is is fair to ...
- 11
1
vote
2
answers
739
views
If, and why torrent traffic is creating greater congestion on a link, than regular traffic of the same size?
For simplicity let's assume that we have very simple LAN. There is some router with uplink to the internet. To this router there is connected some basic 24 ports switch, and from that switch there are ...
- 156
1
vote
1
answer
319
views
How to choose the right congestion control algorithm
In my simulation platform, I noticed three new parameters to me:
Hierarchical Token Bucket (TBF scheduling).
Explicit Congestion Notification (ECN)
Random Early Detection (RED)
When I googled them, ...
- 401
0
votes
2
answers
1k
views
The difference between simple ACK and ACK aggregation improvement in TCP
I came across a new term in TCP ACK aggregation which is a bit new to me. I tried finding a good resource online that precisely explains this concept in a clear but unfortunately all of them are very ...
- 15
1
vote
1
answer
62
views
LSA Weighted Fair Queuing Algorithm
In weighted round robin (weighted fair queuing) scheduling, suppose I have 3 queues (weights):- A (4), B (3) and C (2). I have packets of equal lengths on the 3 queues. Now will the packet output ...
- 323
2
votes
1
answer
3k
views
If we have Fast Retransmit does it mean we have Fast Recovery?
My question is, in which cases is the Fast Recovery activarted ?
Like we have done a Fast Retransmit does it mean that Fast Recovery must be used as well ? For example, we lose a segment and send it ...
- 55