An argument for increasing TCP's initial congestion window

Dukkipati, Nandita; Refice, Tiziana; Cheng, Yuchung; Chu, Jerry; Herbert, Tom; Agarwal, Amit; Jain, Arvind Kumar; Sutin, Natalia

doi:10.1145/1823844.1823848

Cited by 190 publications

(137 citation statements)

References 8 publications

Supporting

Mentioning

132

Contrasting

Unclassified

Order By: Relevance

“…However, MHDs have only been widely adopted recently, and none of those studies explores the network performance of MHDs as in our study. Our study builds upon the rich literature on understanding the behavior of TCPs and content distribution in operational environments (e.g., [19,23,20,15,30,10,4,22]), and our observations confirm some of the findings in those studies.…”

Section: Related Worksupporting

confidence: 82%

See 1 more Smart Citation

Network performance of smart mobile handhelds in a university campus WiFi network

Chen

Jin

Suh

et al. 2012

Proceedings of the 2012 Internet Measurement Conference

View full text Add to dashboard Cite

Smart mobile handheld devices (MHDs) such as smartphones have been used for a wide range of applications. Despite the recent flurry of research on various aspects of smart MHDs, little is known about their network performance in WiFi networks. In this paper, we measure the network performance of smart MHDs inside a university campus WiFi network, and identify the dominant factors that affect the network performance. Specifically, we analyze 2.9TB of data collected over three days by a monitor that is located at a gateway router of the network, and make the following findings:(1) Compared to non-handheld devices (NHDs), MHDs use well provisioned Akamai and Google servers more heavily, which boosts the overall network performance of MHDs. Furthermore, MHD flows, particularly short flows, benefit from the large initial congestion window that has been adopted by Akamai and Google servers. (2) MHDs tend to have larger local delays inside the WiFi network and are more adversely affected by the number of concurrent flows. (3) Earlier versions of Android OS (before 4.X) cannot take advantage of the large initial congestion window adopted by many servers. On the other hand, the large receive window adopted by iOS is not fully utilized by most flows, potentially leading to waste of resources. (4) Some applicationlevel protocols cause inefficient use of network and operating system resources of MHDs in WiFi networks. Our observations provide valuable insights on content distribution, server provisioning, MHD system design, and applicationlevel protocol design.

show abstract

Section: Related Worksupporting

confidence: 82%

“…The behavior of HTTP is complicated. For instance, modern browsers often open multiple TCP connections when browsing a web page [10]. Furthermore, an HTTP flow can send and receive multiple HTTP requests/responses.…”

Section: Performance Metricmentioning

confidence: 99%

Network performance of smart mobile handhelds in a university campus WiFi network

Chen

Jin

Suh

et al. 2012

Proceedings of the 2012 Internet Measurement Conference

View full text Add to dashboard Cite

show abstract

“…The high packet loss evident in some access networks may lead to smaller connections as users decide to discontinue use as they perceive their experience unsatisfactorily. Also, as newer web browser versions open more connections in hope of sustaining higher download speeds [40], the size of connections tends to decrease over time. …”

Section: Discussionmentioning

confidence: 99%

Comparison of User Traffic Characteristics on Mobile-Access versus Fixed-Access Networks

Heikkinen

Berger

2012

Passive and Active Measurement

View full text Add to dashboard Cite

We compare Web traffic characteristics of mobile-versus fixed-access end-hosts, where herein the term "mobile" refers to access via cell towers, using for example the 3G/UMTS standard, and the term "fixed" includes Wi-Fi access.It is well-known that connection speeds are in general slower over mobile-access networks, and also that often there is higher packet loss. We were curious whether this leads mobile-access users to have smaller connections.We examined the distribution of the number of bytesper-connection, and packet loss from a sampling of logs from servers of Akamai Technologies. We obtained 149 million connections, across 57 countries. The mean bytes-per-connection was typically larger for fixed-access: for two-thirds of the countries, it was at least onethird larger. Regarding distributions, we found that the difference between the bytes-per-connection for mobileversus fixed-access, as well as the packet loss, was statistically significant for each of the countries; however the visual difference in plots is typically small. For some countries, mobile-access had the larger connections. As expected, mobile-access often had higher loss than fixed-access, but the reverse pertained for some countries. Typically packet loss increased during the busy period of the day, when mobile-access had a larger increase. Comparing our results from 2010 to those from 2009 of the same time period, we found that connections have become a bit smaller.

show abstract

“…But when it progresses, same method of increments of congestion window leads to exponential increase. Author [18] have suggested to initialize window size to 10 to save 4 RTT times for web traffic. There is a need to develop protocol which follows slow start, but jumps to exponential increase very soon at the start of each iteration and also becomes steady when it's about to reach the next iteration.…”

Section: Proposed Workmentioning

confidence: 99%

“…In our proposed method, we follow the same methodology as mentioned above. M-increase increments congestion window by additive factor P. Here initial value of P is set to 10 as mentioned in [18]. Gradually additive factor is getting decrease by one upon receiving each ack.…”

Section: Proposed Workmentioning

confidence: 99%

TCP M-Start: A New Slow Start Method of TCP to Transfer Data Over Long Fat Pipe Network

Pravinbahi¹,

Pravinchandra²

2017

IJIES

View full text Add to dashboard Cite

Transmission control protocol have gone through various revisions to develop new method of responding to network congestion control as per past, present and future requirements. In cloud computing moving large size of a VM from one data center to another data center over WAN is a challenging task. To regulate data flow, sending of data is carried out in three phases iteratively. Slow start, congestion avoidance and fast recovery. All these methods are functions of ACK reception. Specially in slow start, initially it is very slow but it progresses aggressively as times increases. Because of a slow start, it suffers from low network resource utilization in the high bandwidth delay scenario. At every receipt of ack, the congestion window is either step up by one or a zero. In this paper, efforts are made to improve slow start behavior by changing step up count (scnt) to improve network resource utilization. Experiments and results, observed during the slow start phase, show that throughput has increased to 51.23%, time to reach epoch has reduced to 40%, the position of epoch point has increased to 4 times higher than traditional but it leads to increased 30% of packet drops.

show abstract

An argument for increasing TCP's initial congestion window

Cited by 190 publications

References 8 publications

Network performance of smart mobile handhelds in a university campus WiFi network

Network performance of smart mobile handhelds in a university campus WiFi network

Comparison of User Traffic Characteristics on Mobile-Access versus Fixed-Access Networks

TCP M-Start: A New Slow Start Method of TCP to Transfer Data Over Long Fat Pipe Network

Contact Info

Product

Resources

About