Decentralized compression and predistribution via randomized gossiping

Rabbat,; Haupt,; Singh, Varsha; Nowak, Bernadeta

doi:10.1109/ipsn.2006.244056

Cited by 65 publications

(79 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Junction Tree [13] ensures robust packet delivery for distributed inference, but does not use hybrid hardware. [14], a compressive sensing based approach, simultaneously computes random projections of the sensor data and disseminates them using gossiping. The goal is to reduce communication of Nyquist rate sampled data, as opposed to facilitating subNyquist rate signal sampling for low-power devices.…”

Section: Related Workmentioning

confidence: 99%

RHA: A robust hybrid architecture for information processing in wireless sensor networks

Dang

Bulusu

Feng

et al. 2010

2010 Sixth International Conference on Intelligent Sensors, Sensor Networks and Information Processing

View full text Add to dashboard Cite

Abstract-The paradox of wireless sensor networks (WSNs) is that the low-power, miniaturized sensors that can be deeply embedded in the physical world, are too resource-constrained to capture high frequency phenomena. In this paper, we propose RHA, a robust hybrid architecture for information processing to extend sensing capacity while conserving energy, storage and bandwidth. We evaluate RHA with two real world applications, bio-acoustic monitoring and spatial monitoring. RHA provides accurate signal reconstruction with significant down sampling compared to traditional multi-rate sampling, is robust to noise, enables fast triggering and load balancing across sensors.

show abstract

Section: Related Workmentioning

confidence: 99%

RHA: A robust hybrid architecture for information processing in wireless sensor networks

Dang

Bulusu

Feng

et al. 2010

2010 Sixth International Conference on Intelligent Sensors, Sensor Networks and Information Processing

View full text Add to dashboard Cite

show abstract

“…where g v (x) refers to a subgradient of f v (x), when viewing GGE as an incremental subgradient algorithm 3 . Moreover, the ǫ-averaging time for GGE is bounded above by…”

Section: Gge Convergence Rate: Worst Case Boundmentioning

confidence: 99%

“…3 We explicitly note that this constant is a function of the underlying topology by writing A(G), and A(G) is completely determined by the neighbourhood structure of the network because the maximization is over all x, and for a fixed x, the subgradients are determined by the neighbourhood structure.…”

Section: Gge Convergence Rate: Worst Case Boundmentioning

confidence: 99%

Rates of convergence for greedy gossip with eavesdropping

Üstebay

Oreshkin

Coates

et al. 2008

2008 46th Annual Allerton Conference on Communication, Control, and Computing

View full text Add to dashboard Cite

Abstract-Greedy gossip with eavesdropping (GGE) is a randomized gossip algorithm that exploits the broadcast nature of wireless communications to converge rapidly on grid-like network topologies without requiring that nodes know their geographic locations. When a node decides to gossip, rather than choosing one of its neighbors randomly, it greedily chooses to gossip with the neighbor whose values are most different from its own. We assume that all transmissions are wireless broadcasts so that nodes can keep track of their neighbors' values by eavesdropping on their communications. We have previously proved that GGE converges to the average consensus on connected network topologies. In this paper we study the rate of convergence of GGE, a non-trivial task due to the greedy, data-driven nature of the algorithm. We demonstrate that GGE outperforms standard randomized gossip, and we characterize the rate of convergence in terms of a topologydependent constant analogous to the second-largest eigenvalue characterization for previous randomized gossip algorithms. Simulations demonstrate that the convergence rate of GGE is superior to existing average consensus algorithms such as geographic gossip.

show abstract

“…Recently, random linear coding has also been utilized in sensor networks for failure tolerant aggregation of distributed data [10], [11], [12]. In both Dimakis et al [10] and Rabbat et al [11]'s work, original data is distributed and encoding is only performed on the destination sensor nodes, in order to store them in a coded form.…”

Section: Related Workmentioning

confidence: 99%

“…In both Dimakis et al [10] and Rabbat et al [11]'s work, original data is distributed and encoding is only performed on the destination sensor nodes, in order to store them in a coded form. Compared to Echelon, such uncoded distribution involves larger messaging overhead.…”

Section: Related Workmentioning

confidence: 99%

Echelon: Peer-to-Peer Network Diagnosis with Network Coding

2006

200614th IEEE International Workshop on Quality of Service

View full text Add to dashboard Cite

Abstract-It is critical to monitor the performance and "health" of large-scale peer-to-peer applications. As an example, operators of peer-to-peer live streaming applications may be interested in observing performance bottlenecks, peer failures, and network topologies. In most cases, such observations are used to diagnose potential problems in the protocol design, to troubleshoot network outage, or to improve the Quality of Service of the peer-to-peer network in general. They are not time sensitive in nature, as delayed observations up to minutes or even hours are still valuable. However, such historical and delay-tolerant observations should include measurements of peers that have already failed or departed, as peer dynamics significantly affect the health of peer-to-peer applications. Such a delay-tolerant observation of peer-to-peer applications over a historical period of time is referred to as a diagnosis. In this paper, we present Echelon, a time-insensitive way to construct the diagnosis of a large-scale peer-to-peer application. Replacing the traditional wisdom of logging servers, we leverage the power of network coding to collect application-specific measurements on each peer, and disseminate them to other peers in a coded form. Over time, measurements of departed peers can still be recovered, simply by probing a small subset of peers in the network. Simulation studies have shown that Echelon is highly configurable, bandwidth efficient, and extremely tolerant of peer dynamics, thanks to the advantages of randomized network coding.

show abstract

Decentralized compression and predistribution via randomized gossiping

Cited by 65 publications

References 19 publications

RHA: A robust hybrid architecture for information processing in wireless sensor networks

RHA: A robust hybrid architecture for information processing in wireless sensor networks

Rates of convergence for greedy gossip with eavesdropping

Echelon: Peer-to-Peer Network Diagnosis with Network Coding

Contact Info

Product

Resources

About