Since the diagnosis of poor latency performance of TCP from overbuffering, namely bufferbloat, in cellular networks, there have been a number of proposals to tackle the problem. However, due to the fundamental challenges involved in the goal of persistently achieving the maximum throughput and the minimum latency at the same time, which are in a trade-off relationship, it was not very successful to have a congestion control that performs very closely to its ideal. To this end, we propose a new congestion control, ExLL that extremely restrain its latency to the level of minimum possible RTT (round trip time) while keeping its throughput in the same level of throughput-oriented congestion control schemes such as TCP Cubic.ExLL properly adjusts CWND without increasing RTT excessively or losing throughput even in a mobility scenario. ExLL records around 30% RTT gap while TCP Cubic and BBR shows 419% and 143% gaps.
We raise a question on why the abundant information previously shared between a server and its client is not effectively utilized in the exchange of a new data which may be highly correlated with the shared data. We formulate this question as an encoding problem that is applicable to general data synchronization services including a wide range of Internet services such as cloud data synchronization, web browsing, messaging, and even data streaming. To this problem, we propose a new encoding technique, SyncCoding that maximally replaces subsets of the data to be transmitted with the coordinates pointing to the matching subsets included in the set of relevant shared data, called references. SyncCoding can be easily integrated into a transport layer protocol such as HTTP and enables significant reduction of network traffic. Our experimental evaluations of SyncCoding implemented in Linux shows that it outperforms existing popular encoding techniques, Brotli, LZMA, Deflate, and Deduplication in two practical use networking applications: cloud data sharing and web browsing. The gains of SyncCoding over Brotli, LZMA, Deflate, and Deduplication in the encoded size to be transmitted are shown to be about 12.4%, 20.1%, 29.9%, and 61.2% in the cloud data sharing and about 78.3%, 79.6%, 86.1%, and 92.9% in the web browsing, respectively. The gains of SyncCoding over Brotli, LZMA, and Deflate when Deduplication is applied in advance are about 7.4%, 10.6%, and 17.4% in the cloud data sharing and about 79.4%, 82.0%, and 83.2% in the web browsing, respectively.
[ Overview of the system design and the evaluation scenarios of two use case: 1) Cloud data sharing (left) and 2) Web browsing (right) ]
Prof. Lee sent out a newsletter to FIF on our ongoing research topic. Check this out!
Today, most people use mobile devices for various reasons. We can do many things such as calling, messaging, photographing and gaming through mobile applications. Even now, many people feel frustrated to the fastly decreased battery lifetime by using lots of mobile applications. How can we solve this problem? We find that context-aware application scheduling can be the solution. If we know the next application what we will use by several contexts, we will can preload that application or unload other applications to extend the battery lifetime with the minimal launch delay. We will consider about how to use contexts of users to find the best probability of guessing next application which will be used correctly.
Suppose that a big fire breaks out at a massive shopping complex in which thousands of people are browsing around. People will panic and prominent evacuation exits will soon be overcrowded.
We know by experience that this is not right because such a rush sometimes incurs secondary accidents and also there is no guarantee that those exits are still functional in the ever-chaning environment.
In such an emergency, isn’t it desirable to guide differently how to evacuate according to the locations where groups of people are? For instance, we can imagine a situation where some people are guided to move to the roof, some are guided to stay there, and some others are guided to go to the basement. This obviously sounds plausible, but unfortunately we are incapable of executing this dedicated guidance even with the support of smartphones and high-speed cellular networks, simply because we cannot communicate with the people who their identities are unknown. We are ready to overcome this fundamental limitation of communication with our novel concept of space-time communication.
Prof. Lee sent out a newsletter to FIF on our ongoing research topic, extremely low-latency networking. Check this out!
While being on a public transportation, a few questions sometimes arise. For instance, movement of people is different over time, but why is the operation interval of a public transportation almost static over time?
In a similar manner, why is there no bus route that a lot of people really like to take? Are the public transportation schedule designer aware of the up-to-date transport demand of the entire city?
Is the electronic check-in and check-out system implemented for public transportation efficiently monitoring the transportation volume? We are seeking for the answers to these questions by utilizing smartphones. We are also trying to quantitatively study how much public transportation efficiency can be improved from our answers.