The Effect of Faults on Network Expansion

@article{Bagchi2004TheEO,
  title={The Effect of Faults on Network Expansion},
  author={Amitabha Bagchi and Ankur Bhargava and Amitabh Chaudhary and David Eppstein and Christian Scheideler},
  journal={Theory of Computing Systems},
  year={2004},
  volume={39},
  pages={903-928}
}
We study the problem of how resilient networks are to node faults. Specifically, we investigate the question of how many faults a network can sustain and still contain a large (i.e., linear-sized) connected component with approximately the same expansion as the original fault-free network. We use a pruning technique that culls away those parts of the faulty network that have poor expansion. The faults may occur at random or be caused by an adversary. Our techniques apply in either case. In the… 

Byzantine-Resilient Counting in Networks

Two distributed algorithms are presented that are the first such algorithms that solve Byzantine counting in sparse, bounded degree networks under very general assumptions and can be used for the design of efficient distributed algorithms resilient against Byzantine failures.

Enabling Robust and Efficient Distributed Computation in Dynamic Peer-to-Peer Networks

The main contribution is a randomized distributed protocol that guarantees with high probability the maintenance of a constant degree graph with high expansion even under continuous high adversarial churn.

Storage and search in dynamic peer-to-peer networks

These algorithms are the first-known, fully-distributed storage and search algorithms that provably work under highly dynamic settings (i.e., high churn rates per step) and scalable and do not require any global topological knowledge.

The diameter and connectivity of networks with random dependent faults

This is the first analytic paper which investigates the connectivity and diameter of networks where nodes fail in a random dependent way, and shows that for networks of degree bounded above by a constant Δ, the fault‐free part can not be connected with high probability if p ε 1/O(n1/Δ).

Distributed agreement in dynamic peer-to-peer networks

Towards robust and efficient computation in dynamic peer-to-peer networks

These algorithms are the first-known, fully-distributed, agreement algorithms that work under highly dynamic settings and are localized (i.e., do not require any global topological knowledge), simple, and easy to implement.

Fault-Tolerant Graph Realizations in the Congested Clique

The main result is a O ( f ) -round deterministic algorithm for the degree-sequence realization problem in a n -node Congested Clique, of which f nodes could be faulty ( f < n ); the algorithm uses O ( n 2 ) messages.

Fast byzantine agreement in dynamic networks

We study Byzantine agreement in dynamic networks where topology can change from round to round and nodes can also experience heavy churn (i.e., nodes can join and leave the network continuously over

Local treewidth of random and noisy graphs with applications to stopping contagion in networks

The main parameter of study is k, the number of initially “infected” vertices in the network, and the running time of the algorithms designed for worst-case instances of these edge deletion problems is improved, improving upon the 2 Ω( k ) poly( n ) performance of the best-known algorithms.

Improved Fault Tolerance and Secure Computation on Sparse Networks

This work significantly improves on the above parameters obtaining a protocol with t = αn, d = O(logq n), for constant q > 0 and x = O (t/log n) with the same parameters as above, significantly improving on the bound on the number of left-out parties.

References

SHOWING 1-10 OF 43 REFERENCES

On the fault tolerance of some popular bounded-degree networks

The authors analyze the fault-tolerance properties of several bounded-degree networks and show that an N-node butterfly containing N/sup 1- epsilon / worst-case faults can emulate a fault-free butterfly of the same size with only constant slowdown, becoming the first connected bounded- degree networks known to be able to sustain more than a constant number of worst- case faults.

Load Balancing in Arbitrary Network Topologies with Stochastic Adversarial Input

The long-term (steady state) performance of a simple, randomized, local load balancing technique under a broad range of input conditions is studied and expected and high probability bounds on the total load in the system and the waiting time of jobs in terms of the network topology are developed.

Tolerating linear number of faults in networks of bounded degree

This work answers the question whether almost everywhere agreement can be achieved on bounded degree networks in the presence of up to O(n) faulty nodes (processors) and based on a general technique for simulating on a bounded degree network an algorithm designed for the complete network.

Tolerating a Linear Number of Faults in Networks of Bounded Degree

  • E. Upfal
  • Computer Science
    Inf. Comput.
  • 1994
This work answers the question whether almost everywhere agreement can be achieved on bounded degree networks in the presence of up to O ( n ) faulty nodes (processors) in the affirmative and proposes a general technique for simulating on a bounded degree network an algorithm designed for the complete network.

Reconfiguring Arrays with Faults Part I: Worst-Case Faults

If faulty nodes are allowed to communicate, but not compute, then an N-node one-dimensional array can tolerate $\log^k N$ worst-case faults, for any constant $k > 0$, and still emulate a fault-free array with constant slowdown, and this bound is tight.

Stability of load balancing algorithms in dynamic adversarial systems

The proof techniques are very different from the ones used by Muthukrishnan and Rajaraman, and yield a simpler proof and tighter bounds on the difference in loads, and apply to the problem of routing packets in adversarial systems.

Routing complexity of faulty networks

It is shown that an oracle access to links (as opposed to local routing) may reduce significantly the complexity of the routing problem, and tight upper and lower bounds for the failure probability are provided, which permits routing both for the hypercube and for the d‐dimensional mesh.

Fault tolerance in networks of bounded degree

It is believed that in forseeable technologies the number of faults will grow with the size of the network while the degree will remain practically fixed, and the question whether it is possible to avoid the connectivity requirements by slightly lowering the authors' expectations is raised.

On the fault tolerance of the butterfly

There is a O-1 law on the existence of a linearized component of the butterfly network such that for p above p*, the faulted butterfly almost surely contains a linear-sized component, whereas for p below px, the Faulted Butterfly almost surely does not contain a linearised component.

Reconfiguring a hypercube in the presence of faults

Algorithms for embedding an N/2-node hypercube in an N- node hypercube with faulty processors and ways to produce embeddings which allow for low delay simulations, as well as ways to use a faulty hypercube to efficiently simulate a completely functioning hypercube of the same size are described.