(→TCP Congestion Control Algorithms over a live HSDPA Network) |
(→Results of the experiments) |
||
Riga 30: | Riga 30: | ||
* In the case of the uplink channel the algorithms exhibited similar performances, due to the fact that the moderate size of cwnd forced TCP Cubic and TCP Bic to use the linear probing in order to be TCP friendly (see [http://netsrv.csc.ncsu.edu/export/cubic_a_new_tcp_2008.pdf [1]]). | * In the case of the uplink channel the algorithms exhibited similar performances, due to the fact that the moderate size of cwnd forced TCP Cubic and TCP Bic to use the linear probing in order to be TCP friendly (see [http://netsrv.csc.ncsu.edu/export/cubic_a_new_tcp_2008.pdf [1]]). | ||
− | ==Results of the experiments== | + | ==Results of the experiments and Conclusions== |
We report cumulative distribution function plots: | We report cumulative distribution function plots: | ||
* [[Tcp_over_Hsdpa:goodput_downlink | downlink]] and [[Tcp_over_Hsdpa:goodput_uplink | uplink]] channel goodput CDFs | * [[Tcp_over_Hsdpa:goodput_downlink | downlink]] and [[Tcp_over_Hsdpa:goodput_uplink | uplink]] channel goodput CDFs | ||
Riga 36: | Riga 36: | ||
* Goodput-RTT plots for [[Tcp_over_Hsdpa:goodput_rtt_downlink|downlink]] and [[Tcp_over_Hsdpa:goodput_rtt_uplink|uplink]] channels. | * Goodput-RTT plots for [[Tcp_over_Hsdpa:goodput_rtt_downlink|downlink]] and [[Tcp_over_Hsdpa:goodput_rtt_uplink|uplink]] channels. | ||
* Timeouts and packet loss ratio bar graphs for [[Tcp_over_Hsdpa:plrto_downlink|downlink]] and [[Tcp_over_Hsdpa:plrto_rtt_downlink|uplink]] channels (3 times timeouts and 2 times packet loss ratio of BIC/CUBIC wrt NewReno/Westwood+) | * Timeouts and packet loss ratio bar graphs for [[Tcp_over_Hsdpa:plrto_downlink|downlink]] and [[Tcp_over_Hsdpa:plrto_rtt_downlink|uplink]] channels (3 times timeouts and 2 times packet loss ratio of BIC/CUBIC wrt NewReno/Westwood+) | ||
+ | |||
+ | Abstract | ||
+ | TCP BIC/CUbic is the default cc in Linux. As consequence roughly 50% of servers employs TCP Bic/Cubic []. | ||
+ | In this paper we show the reasons not to deploy TCP Bic/Cubic, which are rooted in its more aggressive probing phase. In particular, in common network conditions, TCP BIC/CUBIC exhibits: 1. a larger RTT average wrt to TCP NewReno or TCP Westwood+; 2. a larger number of retransmission wrt to TCP NewReno or TCP Westwood+; 3 larger throughput but same goodput wrt toTCP NewReno or Westwood+. | ||
+ | In other terms, its more aggressive probing increases both throughput and retransmission but leaving unchanged the goodput that is neutral for the user but negative for the network. | ||
==Congestion window and RTT evolutions== | ==Congestion window and RTT evolutions== |
Goodput | RTT | PLR/Timeouts | Goodput/RTT | |
Back | downlink/ uplink | downlink/ uplink | downlink / uplink | downlink / uplink |
We report results obtained by an extensive measurement campaign over a live HSDPA network involving more than 3000 flows resulting in around 60 hours of active measurements. The considered TCP variants are the following:
The experimental evaluation has been carried out by accessing the public Internet using a commercial HSDPA card (see figure below) and by employing the linux Kernel 2.6.24.
The following instantaneous variable measurements have been collected by using the libnetmeas library:
The main findings of the investigation can be summarized as follows:
We report cumulative distribution function plots:
Abstract TCP BIC/CUbic is the default cc in Linux. As consequence roughly 50% of servers employs TCP Bic/Cubic []. In this paper we show the reasons not to deploy TCP Bic/Cubic, which are rooted in its more aggressive probing phase. In particular, in common network conditions, TCP BIC/CUBIC exhibits: 1. a larger RTT average wrt to TCP NewReno or TCP Westwood+; 2. a larger number of retransmission wrt to TCP NewReno or TCP Westwood+; 3 larger throughput but same goodput wrt toTCP NewReno or Westwood+. In other terms, its more aggressive probing increases both throughput and retransmission but leaving unchanged the goodput that is neutral for the user but negative for the network.
Typical congestion window and RTT evolutions of the considered TCPs in the case of a single flow over the HSDPA downlink
Goodput | RTT | PLR/Timeouts | Goodput/RTT | |
Back | downlink/ uplink | downlink/ uplink | downlink / uplink | downlink / uplink |
We report results obtained by an extensive measurement campaign over a live HSDPA network involving more than 3000 flows resulting in around 60 hours of active measurements. The considered TCP variants are the following:
The experimental evaluation has been carried out by accessing the public Internet using a commercial HSDPA card (see figure below) and by employing the linux Kernel 2.6.24.
The following instantaneous variable measurements have been collected by using the libnetmeas library:
The main findings of the investigation can be summarized as follows:
We report cumulative distribution function plots:
Abstract TCP BIC/CUbic is the default cc in Linux. As consequence roughly 50% of servers employs TCP Bic/Cubic []. In this paper we show the reasons not to deploy TCP Bic/Cubic, which are rooted in its more aggressive probing phase. In particular, in common network conditions, TCP BIC/CUBIC exhibits: 1. a larger RTT average wrt to TCP NewReno or TCP Westwood+; 2. a larger number of retransmission wrt to TCP NewReno or TCP Westwood+; 3 larger throughput but same goodput wrt toTCP NewReno or Westwood+. In other terms, its more aggressive probing increases both throughput and retransmission but leaving unchanged the goodput that is neutral for the user but negative for the network.
Typical congestion window and RTT evolutions of the considered TCPs in the case of a single flow over the HSDPA downlink