Lev B. Sofman

Analytical Model for Hierarchical Cache Optimization in IPTV Network

In an IPTV network, video on demand and other video services generate large amount of unicast traffic from video hub office (VHO) to subscribers and, therefore, require additional bandwidth and equipment resources in the network. To reduce this traffic and overall network cost, a portion of the video content (the most popular titles) may be stored in caches closer to subscribers, e.g., in a digital subscriber line access multiplexer (DSLAM), a central office (CO), or in an intermediate office (IO). The problem becomes to minimize cost by optimizing cache memory placement and amount. We present an analytical model of hierarchical cache optimization. This model depends on several basic parameters: traffic volume, cache hit rate as a function of memory size, topology (i.e. number of DSLAMs, service routers at CO, service switches at IO locations), and cost parameters. Some reasonable assumptions about a network cost structure and a hit rate function allow us to obtain an analytically optimal solution for the problem. We then analyze the factors that affect this solution.

Dimensioning of active broadcast channels in access IPTV network

The number of active broadcast TV channels between central office and access node and its sensitivity to channel popularity distribution and to the number of set top boxes simultaneously tuned to broadcast channels is investigated.

Comparative Account of Bandwidth Efficiency in Optical Burst Switching and Optical Circuit Switching Networks

The performance of Optical Burst Switched (OBS) networks is compared to that of Optical Circuit Switched (OCS) networks in terms of bandwidth efficiency. Two traffic flow models are employed to investigate the effects of traffic grooming, packetization, buffering and burst assembly on bandwidth efficiency and utilization. Traffic is represented by fluid-flow continuous bit streams in the first model and by packet-based discrete flows in the second model. It is shown that the parameters of traffic characteristics and burst assembly have significant impact on the relative bandwidth savings of OBS. The bandwidth efficiency of OBS compared to OCS should not be taken for granted and has to be examined carefully on a case by case basis.

U-turn caching optimization in IPTV networks

In an IPTV network, a large amount of unicast traffic from the Video Head Office (VHO) to subscribers requires significant BW and equipment resources in the network. Caching of video content allows reduction of bandwidth and IPTV network cost. In this paper, we consider a caching architecture in which part of the content is cached at the edge of the network (e.g., at DSLAMs), and the remaining part is stored in server(s) at the VHO. We assume that the community of subscribers subtended at each cache location (DSLAM) may have unique interests. Therefore, the list of cached items is customized for each cache location (e.g., individual DSLAM), and different cache locations should store different content. We consider U-turn caching where titles cached at a given DSLAM may serve the requests of subscribers that are subtended not only at this DSLAM but also from those subtended at neighboring DSLAMs. We discuss how the U-turn caching architecture allows increased cache effectiveness (hit rate) by sharing cache resources in the access nodes (DSLAMs).

Enhancement of bandwidth efficiency by traffic grooming in optical-cross-connect based networks

It is anticipated that future core networks may evolve to Optical Cross-Connect (OXC) based architectures where switching is carried out at the granularity of wavelength-based circuits. While the switch fabric is an important component of an OXC, the number of ports is typically a more significant contributor to its cost. Reducing the number of ports (or number of wavelengths per link) will also result in lower transmission-systems costs. We investigate the effect of incorporating some packet switching functionality in the OXC and try to quantify the bandwidth savings that can be gained by doing so, under certain assumptions about the network and the traffic characteristcs. Results show that taffic grooming and statistical multiplexing gain, which are introduced by the packet handling functionality, lead to better channel utilization. This can result in lowering the number of channels and ports and in reducing the OXC size and cost.

Effect of traffic variability on statistical multiplexing gain in a bufferless continuous traffic model

The effect of traffic variability on statistical multiplexing gain is analyzed in a bufferless continuous fluid-flow traffic model. Two different methods are used to account for traffic loss, namely a method based on overflow probability and another one based on traffic loss ratio. It is shown that for both methods, bandwidth savings due to statistical multiplexing gain (SMG) can be significant and increases with increasing traffic variability. It is also shown that SMG and channel utilization increase as the number of composite traffic streams increases and as traffic loss probability/ratio is lowered.