Timely and accurate studies on the composition and nature of the Internet are crucial for continued research and innovation. The aim of this research is to aid and service emulated and simulated research methods where realism is dependent on the accuracy of Internet statistics. This study has captured anonymised, and analysed Internet traffic entering and exiting a university network. Passive measurement techniques are used to discover the mean and distribution of packet sizes. A range of Network and Transport layer statistics, such as DCSP, SACK, ECN, MSS usage, Window Scaling and Timestamps are also investigated. Active measurement techniques are used to determine Round Trip Times (RTTs) and the number of lost and misordered packets. Prior work is leveraged to explore changes in the composition and nature of the Internet over the last 10 years.
The proliferation of Wireless LANs and the increasing integration of voice into data networks has created the potential for VoWLANs (Voice over Wireless Local Area Networks). This technology has immense cost saving potential and the ability to provide better service and functionality. However, before the integration of VoWLAN is possible, handoff delays must be reduced. Currently, the connectivity transition that occurs from moving between APs (Access Points) is too long, causing poor voice quality and call dropouts. An experimental approach is used to investigate a particular handoff delay known as the scanning delay. The study concludes that the primary source of delay in the scanning process is caused by overlapping channels in the 2.4GHz band.
This study experimentally compares the performance of three different multi hop ad hoc network routing protocols. Traditional routing protocols have proven inadequate in wireless ad hoc networks, motivating the need for ad hoc specific routing protocols. This study tests link state, distance vector and biologically inspired approaches to routing using OLSR, Babel and BATMAN routing protocols. The importance of OSI layers is also discussed. This study concludes that the routing protocol's overhead is the largest determinant of performance in small multi hop ad hoc networks. The results show that Babel outperforms OLSR and BATMAN routing protocols and that the OSI layer of the routing protocol has little impact on performance.
Ethernet data rates have increased many orders of magnitudes since standardisation in 1982. Despite these continual data rates increases, the 1500 byte Maximum Transmission Unit (MTU) of Ethernet remains unchanged. Experiments with varying latencies, loss rates and transaction lengths are performed to investigate the potential benefits of Jumboframes on the Internet. This study reveals that large MTUs offer throughputs much larger than a simplistic overhead analysis might suggest. The reasons for these higher throughputs are explored and discussed.
Packet corruption negatively affects TCP through-puts because losses are interpreted as congestion. To enable TCP to operate effectively over wireless networks, a number of reliability techniques such as Forward Error Correction (FEC) and Automatic Repeat Request (ARQ) are used. These provide reliability at the expense of increased overheads. This study provides experimental results of a new distributed Performance Enhancing Proxy (PEP) called D-Proxy. This proxy can provide reliability to wireless links with minimal overhead. The results show that D-Proxy can provide near-optimal performance in the presence of high loss rates. It is suggested that D-Proxy could be used to replace current ARQ mechanisms.
The founding idea behind this study was that 802.11 acks and TCP acks are substantial contributors to 802.11 over-heads, yet, they both provide the same functionality; reliability. Initial experiments suggest that 802.11 acks contribute to over 20% of the overhead in 802.11 networks. Unfortunately, without 802.11 acks, paths with RTTs greater than a millisecond are unable to utilise this additional performance because lost packets, which occur frequently in unacknowledged (NoAck) 802.11, are interpreted as congestion. This study experiments with a range of PEPs (Performance Enhancing Proxies) which retransmit lost packets. A new proxy, known as D-Proxy, designed to solve the shortcomings of previous I-TCP and Snoop proxies, is experimentally developed and tested in Linux. D-Proxy is a distributed, proactive proxy that caches, analyses and resends packets based on TCP sequence numbers. The results suggest that D-Proxy can substantially improve 802.11 throughputs.