Limits...
Relay discovery and selection for large-scale P2P streaming

View Article: PubMed Central - PubMed

ABSTRACT

In peer-to-peer networks, application relays have been commonly used to provide various networking services. The service performance often improves significantly if a relay is selected appropriately based on its network location. In this paper, we studied the location-aware relay discovery and selection problem for large-scale P2P streaming networks. In these large-scale and dynamic overlays, it incurs significant communication and computation cost to discover a sufficiently large relay candidate set and further to select one relay with good performance. The network location can be measured directly or indirectly with the tradeoffs between timeliness, overhead and accuracy. Based on a measurement study and the associated error analysis, we demonstrate that indirect measurements, such as King and Internet Coordinate Systems (ICS), can only achieve a coarse estimation of peers’ network location and those methods based on pure indirect measurements cannot lead to a good relay selection. We also demonstrate that there exists significant error amplification of the commonly used “best-out-of-K” selection methodology using three RTT data sets publicly available. We propose a two-phase approach to achieve efficient relay discovery and accurate relay selection. Indirect measurements are used to narrow down a small number of high-quality relay candidates and the final relay selection is refined based on direct probing. This two-phase approach enjoys an efficient implementation using the Distributed-Hash-Table (DHT). When the DHT is constructed, the node keys carry the location information and they are generated scalably using indirect measurements, such as the ICS coordinates. The relay discovery is achieved efficiently utilizing the DHT-based search. We evaluated various aspects of this DHT-based approach, including the DHT indexing procedure, key generation under peer churn and message costs.

No MeSH data available.


The cumulative distribution of the error ratios of the RTT measurements.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5391927&req=5

pone.0175360.g003: The cumulative distribution of the error ratios of the RTT measurements.

Mentions: To evaluate the performance of RTTometer, we randomly choose about 8600 pairs in Table 2. For these pairs, we did not conduct any pre-process on the measurement results; instead, we take all the available results into account, 6 measurements for each. For each pair, we compute the error ratio of each RTTometer measurement with the median value,(RTT*-RTTmedian)/RTTmedian,where RTT* is each measurement data point. In Fig 3(a) we plot the CDF of this error ratio. We observe that the error ratios closely cluster at the vicinity of zero. This indicates that the RTTometer measurement is quite accurate. The median of the RTTometer measurements can serve as a reasonably good estimate of the real RTT value of one pair. To evaluate the accuracy of the King measurements, for each pair, we take 10 measurements and compute the error ratio of each King measurement as (RTTKing − RTTRTTometer)/RTTRTTometer, in which RTTRTTometer is the median of the RTTometer RTTs of the pair. Fig 3(b) plots this error ratio for all 74, 339 pairs with RTT between 5 msec and 2000 msec which appear in both the RTTometer and King results. As shown in Fig 3(b), the King RTTs exhibit some variation. It is an indication of the incurred measurement errors in King more severe than those in RTTometer. These errors inevitably degrade the performance of the ICS-based relay discovery and selection since the previous ICS-coordinates are computed based on the King RTTs.


Relay discovery and selection for large-scale P2P streaming
The cumulative distribution of the error ratios of the RTT measurements.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5391927&req=5

pone.0175360.g003: The cumulative distribution of the error ratios of the RTT measurements.
Mentions: To evaluate the performance of RTTometer, we randomly choose about 8600 pairs in Table 2. For these pairs, we did not conduct any pre-process on the measurement results; instead, we take all the available results into account, 6 measurements for each. For each pair, we compute the error ratio of each RTTometer measurement with the median value,(RTT*-RTTmedian)/RTTmedian,where RTT* is each measurement data point. In Fig 3(a) we plot the CDF of this error ratio. We observe that the error ratios closely cluster at the vicinity of zero. This indicates that the RTTometer measurement is quite accurate. The median of the RTTometer measurements can serve as a reasonably good estimate of the real RTT value of one pair. To evaluate the accuracy of the King measurements, for each pair, we take 10 measurements and compute the error ratio of each King measurement as (RTTKing − RTTRTTometer)/RTTRTTometer, in which RTTRTTometer is the median of the RTTometer RTTs of the pair. Fig 3(b) plots this error ratio for all 74, 339 pairs with RTT between 5 msec and 2000 msec which appear in both the RTTometer and King results. As shown in Fig 3(b), the King RTTs exhibit some variation. It is an indication of the incurred measurement errors in King more severe than those in RTTometer. These errors inevitably degrade the performance of the ICS-based relay discovery and selection since the previous ICS-coordinates are computed based on the King RTTs.

View Article: PubMed Central - PubMed

ABSTRACT

In peer-to-peer networks, application relays have been commonly used to provide various networking services. The service performance often improves significantly if a relay is selected appropriately based on its network location. In this paper, we studied the location-aware relay discovery and selection problem for large-scale P2P streaming networks. In these large-scale and dynamic overlays, it incurs significant communication and computation cost to discover a sufficiently large relay candidate set and further to select one relay with good performance. The network location can be measured directly or indirectly with the tradeoffs between timeliness, overhead and accuracy. Based on a measurement study and the associated error analysis, we demonstrate that indirect measurements, such as King and Internet Coordinate Systems (ICS), can only achieve a coarse estimation of peers’ network location and those methods based on pure indirect measurements cannot lead to a good relay selection. We also demonstrate that there exists significant error amplification of the commonly used “best-out-of-K” selection methodology using three RTT data sets publicly available. We propose a two-phase approach to achieve efficient relay discovery and accurate relay selection. Indirect measurements are used to narrow down a small number of high-quality relay candidates and the final relay selection is refined based on direct probing. This two-phase approach enjoys an efficient implementation using the Distributed-Hash-Table (DHT). When the DHT is constructed, the node keys carry the location information and they are generated scalably using indirect measurements, such as the ICS coordinates. The relay discovery is achieved efficiently utilizing the DHT-based search. We evaluated various aspects of this DHT-based approach, including the DHT indexing procedure, key generation under peer churn and message costs.

No MeSH data available.