cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1805
Views
5
Helpful
5
Replies

MSS and Bandwidth-delay Product

alihassan1618
Level 1
Level 1

Which value determines the amount of traffic that a network path can hold in transit?

MSS or Bandwidth-delay product???

 

I think BPD do the job because as far as i know about maximum segment size is that it is the value presented by your system. You can change this value to tune TCP performance by allowing your system to receive the largest possible segments without Fragmentation (but does this effect on transit nodes?). <--- i'm confused here!

on other hand, Bandwidth-delay product actually determines the amount of data that can be in transit in the network. It is the product of the available bandwidth and the latency, or RTT. BDP is a very important concept in a Window based protocol such as TCP.

1 Accepted Solution

Accepted Solutions

Hello AliHassan1618,

in an exam question at CCNA or CCNP level choicing the answer BDP should be enough.

Because half the BDP is still related to BDP, isn't it ?

 

MSS is a different concept that tells how big can be the L4 PDU and it is negotiated during TCP session setup.

 

Hope to help

Giuseppe

 

View solution in original post

5 Replies 5

Giuseppe Larosa
Hall of Fame
Hall of Fame

Hello Alihassan1618,

the bandwidth delay product provides you the upper limit of performance you can achieve in a network path.

Please notice that delay should be the two way delay RTT, RTT is two way delay Round Trip time. Latency is one way delay they are not the same thing.

in BDP we should use RTT time as we need to take in account the time for a TCP Ack to reach the original sender.

see

https://en.wikipedia.org/wiki/Bandwidth-delay_product

 

MSS has to be set to the max value that is supported on all the links on the path to avoid fragmentation that would impact on performance.

 

TCP extended window size / window scale  concept has been introduced for links with high speed and high two way delay.

see

https://tools.ietf.org/html/rfc7323

 

Hope to help

Giuseppe

 

Joseph W. Doherty
Hall of Fame
Hall of Fame
The answer, to your first question, is one half of BDP of the overall path. NB: to achieve this using TCP, the receiving host's RWIN must be the BDP. The Wiki reference that Giuseppe provides isn't fully correct/clear in stating "In data communications, the bandwidth-delay product is the product of a data link's capacity (in bits per second) and its round-trip delay time (in seconds).[1] The result, an amount of data measured in bits (or bytes), is equivalent to the maximum amount of data on the network circuit at any given time, i.e., data that has been transmitted but not yet acknowledged. " What happens is half of the BDP volume has actually been received and ACKed by the receiving host, but the second half is in physical transit before the sender actually receives the ACK. So the network path only holds half of the BDP at anyone time, but you must size for the BDP to allow time for the ACKs to be received. (Actually with TCP you also need to also size for BDP plus the bit size times to allow for receiving the first two packets.)

MSS would help control how efficient your traffic is, i.e. useful data vs. various protocol overheads. The larger the MSS, the less protocol overhead, as a percentage of bandwidth, due to L2, L3 and L4 protocols, until you reach MTU, then it becomes a bit more complicated. (If you're wondering doesn't L3 fragmentation always increase overhead, well it might or might not. The L4 overhead in reduced but unsure the L3 fragmentation overhead is more than would be seen in same sized packets where MSS doesn't need to span more than one. However, often with fragmentation a MSS segment might be in a max MTU packet followed by a smaller IP packet and then the L2 and L3 overhead would be more. For example, if you sized MSS to take up exactly some of number of MTU sized packets, the overhead might be a tad less.)

Regarding MSS and "normal" transit nodes, they generally don't care about L4 (excluding L4 ACL ACEs, etc.)

Also BTW, you seldom "tune" MSS except in cases to reduce it when you know setting it to a certain size will cause IP fragmentation somewhere along the network path.

Hi Joseph W. Doherty,  Thanks for the reply and brief answer. but there is no option availabe to choost "Half of BDP" so what is the correct answer or near to the correct answer. 

Question still remain unanswered...  Which value determines the amount of traffic that a network path can hold in transit? BDP or MSS??? 

Hello AliHassan1618,

in an exam question at CCNA or CCNP level choicing the answer BDP should be enough.

Because half the BDP is still related to BDP, isn't it ?

 

MSS is a different concept that tells how big can be the L4 PDU and it is negotiated during TCP session setup.

 

Hope to help

Giuseppe

 

BDP (at least for TCP) set on the receiving host's RWIN does determine the amount of data the network circuit can hold. However, the Wiki BDP reference, again, states that a BDP amount of data is actually what would be in the network path, and that shouldn't be the case. Sorry my "correction" to the Wiki BDP reference confused you as whether BDP or MSS was the controlling value and my additional information on MSS noting that it has some control over protocol overhead probably also confused you too.
Review Cisco Networking products for a $25 gift card