ALTQ

Abstract: We discuss the syntax and semantics of Japanese alternative questions (AltQs), and argue that they are underlyingly disjoined polar questions (PolQs). The argument is based on the following three observations: (i) the minimum syntactic size of a disjunct in a Japanese AltQ is VP or clausal, (ii) sentential operators above the disjunction in an AltQ are never interpreted in both disjuncts, and (iii) the clause-final particle in an AltQ has to be matched with the particle at the end of the first disjunct. It will be shown that these facts receive straightforward explanations under the analysis that treats Japanese AltQs as disjoined PolQs while they are problematic in a scoping analysis. We also propose a compositional semantics associated with the analysis, and address a potential problem regarding the intervention effect.

Abstract: Class based queueing (CBQ) is a link-sharing and resource management mechanism for packet networks. We have ported the CBQ implementation on FreeBSD, also known as alternate queueing (ALTQ), to the Linux platform. This paper investigates the performance of CBQ in terms of latency, bandwidth guarantee and delay. CBQ is found to be able to preserve bandwidth allocated to the classes. However it does not scale well with the number of data flows as it incurs additional delay with each new data flow admitted. Due to this additional delay, the admission control cannot predict the delay experienced by the existing flows after admitting a new flow. Although CBQ is able to provide lower delay to higher priority flows, it also introduces jitter to the traffic owing to its packet scheduling artifact. As such, CBQ is not suitable for fine-grained scheduling of real-time traffic. It is more suited for providing bandwidth guarantee to aggregated data flows.

Abstract: This paper presents a study carried out on a weighted fair queuing implementation for Unix routers-the WFQ implementation of the ALTQ project. It shows the WFQ/ALTQ weaknesses and explains why we cannot expect an interesting behavior from a system using such a scheduler. The conclusions here presented are supported by a set of tests using UDP traffic only. With a tool developed in our laboratory, we were able to show that changing the classes' weights does not necessarily result on a different quality of service for each of the existing classes. To achieve this differentiation, the lengths of the queues which serve the scheduler (one for each class) must be increased beyond reasonable values. We found that the low-level dynamics of FreeBSD systems practically turns WFQ schedulers useless. The same is applicable to any other work-conserving discipline. Thus, an important conclusion of this paper, is that one must design very carefully the platforms that support work conserving disciplines in order to expect adequate behaviors from those systems, in terms of QoS provision.

Abstract: At the Laboratory of Communications and Telematics of the University of Coimbra we have been working on the IP QoS field. Our work includes researching, testing and implementing mechanisms that can be used on current IP networks in order to make them able to provide different performance levels to traffic with different QoS needs. Our work is being supported by a testbed of INTEL hosts running the FreeBSD operating system patched with ALTQ software. This is a common platform in this research field. We chose it, in part, for that reason - there is much work produced for this environment in which we are interested - and also because it is fully open, allowing us to freely test new approaches for QoS provision. Despite the amount of work being done in this area, soon we realized the lack of a tool to monitor with detail the operation of the IP layer on routers, in particular, the mechanisms related to QoS provision. It was evident we needed such a tool, so we decided to implement one - which we named QoStat. This paper presents QoStat a GUI based tool, which gives us the possibility to visualize in real time the most important values related to QoS provision. With it we can also change on the fly the most important operational parameters of the mechanisms associated with QoS provision. In this way, it is possible to understand with depth the behavior of those systems - for instance, to study cause-effect relations between the values of the referred parameters and the QoS in fact provided to the different IP traffic flows or classes.