The proliferation of diverse network technologies has enabled concurrent utilization of multiple network media, such as WiFi and 4G/5G links on modern smartphones. While this advancement facilitates simultaneous downloading of varied network contents, traditional multi-connection management methods, like reception-driven requesting schemes, often struggle to adapt to dynamic network conditions. This paper introduces a novel timer-driven requesting scheme applied to both parallel TCP and parallel MPTCP (Multipath TCP) methods, addressing the limitations of existing approaches. Unlike previous studies that relied solely on simulations, we present an experimental evaluation of this scheme on real machines. Our extensive experimentation reveals significant advantages of the timer-driven requesting scheme, particularly its superior adaptability in fluctuating network environments. These findings not only demonstrate the scheme’s practical viability but also underscore its potential to revolutionize content retrieval in multi-network scenarios, paving the way for more efficient and resilient data transfer mechanisms in increasingly complex network ecosystems.

"Parallel downloading" is one of the most effective technologies which speed up downloading large files on the Internet. As ubiquitous networks will be realized in the near future, downloading large files quickly is also very important on various types of networks. However, if we use the parallel downloading method on dynamic networks, such as ad hoc networks, it will not work well because of high loss rate and delay. To address this issue, we adopt UDP as a transport layer protocol instead of TCP which guarantees perfect in-order delivery unnecessary for file downloading. In our previous paper, a prototype of the parallel downloading method using HTTP over UDP is developed, and it is confirmed that our method outperforms the traditional method which adopts TCP, especially when loss rate is high and delay is large. In this paper, the key feature of the parallel downloading method using HTTP/UDP is evaluated through simple theoretical estimation and some simulation experiments. As a result, it is confirmed that the large division number tends to yield short download time. The proposed method can be an attractive protocol to distribute large files on dynamic networks composed of wireless links.

When ubiquitous networks pervade in future, applications such as file transfer and Web browsing will be used also on ad-hoc networks. End-to-end reliable communication is necessary for these applications, and TCP has usually been used as a transport protocol. However, it is known that the performance of fairness and transfer rate decreases when TCP is used on ad hoc networks. Particularly, multi-connection flows of TCP have danger of unfair communication among all flows sharing a wireless link. Then, we pay attention to the multistream mechanism of SCTP to improve unfair communication of TCP multi-connection. SCTP can manage multiple data streams in one association by using the multi-stream mechanism. The performance of reliable transport protocols such as TCP and SCTP easily suffers from frequent route failures and contention on the wireless channel. To improve the problem for mobile ad hoc networks, we propose a scheme which adds the mechanism of Fixed RTO and dynamic delayed SACK to SCTP (shortly, the proposal). And, we evaluated the proposal by a network simulator QualNet.

"Parallel downloading" is one of the most effective technologies which speed up downloading large files on the Internet. As ubiquitous networks will be realized in the near future, downloading large files quickly is also very important on various types of networks. However, if we use the parallel downloading method on dynamic networks, such as ad hoc networks, it will not work well because of high loss rate and delay. To address this issue, we adopt UDP as a transport layer protocol instead of TCP which guarantees perfect in-order delivery unnecessary for file downloading. We have implemented a prototype of the parallel downloading method using HTTP over UDP, and it is confirmed that our method outperforms the traditional method which adopts TCP, especially when loss rate is high and delay is large. Our proposal will be effective on dynamic networks such as ad hoc networks included in the ubiquitous network environments

While traditional downloading programs usually select one mirror server for retrieving a file, it is more effective to select several mirror servers and to fetch different parts of the same file from the servers in parallel from the viewpoints of downloading time and fault-tolerance of network service. We have already proposed an adaptive parallel downloading method that enables us to stabilize and accelerate the downloading time even if the network offers a wide variety of bandwidths. This paper describes a new adaptive block requesting method which improves our original method. The new method has the following features: (1) suitable parameters based on a theoretical model and (2) introduction of changing servers. Practical evaluations that focus on downloading speed and stability of downloading time confirm that the improved method is more effective than the original method.

As a variety of communication devices emerging, we can easily install two or more communication interfaces on a PC or a PDA and make a multi-home environment. To utilize this multi-home environment effectively, SCTP (stream control transmission protocol) has been proposed and standardized. Extending functions of the original SCTP, a strategy which can adapt SCTP to real-time communications has been proposed. However, the typical existing strategy has not enough considered the switching conditions from the viewpoint of transmission delay, so there is room to improve them. In this paper, we clarify the drawbacks of the existing strategy and their influences on the performance. Then, we propose a new strategy of path switching to reduce the delay and to increase the amount of data successfully delivered. The effect of the proposed strategy is discussed and confirmed through experimental evaluations.

As the Internet users increase, asymmetric networks which provide asymmetric bandwidth or delay for the uplink and downlink have become a great attraction. However asymmetric networks which use both terrestrial and satellite links at the same time have not been adequately investigated. Therefore, this paper proposes a new formula for TCP performance evaluation for the asymmetric networks. Using this evaluation formula, we calculate the throughput of TCP Reno over the asymmetric networks taking slow start into account. The calculation results are compared with the following: (1) the value based on an existing theoretical formula, (2) the outputs of simulation by NS (Network Simulator), and (3) the experimental results using VSAT (Very Small Aperture Terminal) satellite communication system for satellite links and the Internet for terrestrial links. As a result, it is shown that the new formula is more precise than the one already proposed.

Since the traffic of NetNews is increasing, storage of all articles becomes a serious problem from the viewpoint of wasting network bandwidth and the amount of disk usage. In addition, not all incoming articles are read by users. We have proposed several caching algorithms to overcome this problem and shown that a selective prefetch scheme gives the best system performance among the proposed ones. However, since the selective prefetch scheme employed a simple prefetching policy, the scheme gave a low hit ratio in some cases. Therefore, the paper improves the selective prefetch scheme from the viewpoint of the amount of disk usage as well as hit ratio. We divide the scheme into three factors: reference span, criterion, and threshold in criterion. Through simulation experiments using actual NetNews logs, we investigate the influence of the factors of the reference span and the threshold to system performance. As a result, it is shown that the reference span is a more significant factor than the threshold and when the selective prefetch scheme has a value of around seven days, reference span keeps a high hit ratio and reduces the amount of disk usage.