A new analysis of the false positive rate of a Bloom filter

Christensen, Ken; Roginsky, Allen; Jimeno, Miguel

doi:10.1016/j.ipl.2010.07.024

Cited by 95 publications

(62 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We acknowledge that there might be more accurate techniques for computing the false positive rates (e.g., [15]); our findings, however, show that the difference in false positives resulting from [15] and [24] was negligible and did not affect our results. We therefore elected to rely on the estimation of P f which appears in [24].…”

Section: Introductionmentioning

confidence: 80%

“…Bloom Filters: Bloom filters have been first proposed by Bloom in 1970; we refer the readers to [15] for detailed information on Bloom filters. In SPV clients [2], a Bloom filter B of an SPV client is specified by the maximum number of elements that it can fit, denoted by M , without exceeding its target false-positive rate Pt.…”

Section: Introductionmentioning

confidence: 99%

“…Similarly, to query the presence of an element x ∈ {0, 1} * in B, then this entails computing ∧ Bloom filters can generate a number of false positives, but cannot result in false negatives. The literature features a number of techniques to estimate the false positive rate of Bloom filters [15,24]. In this paper, we focus on the proposal due to Mullin et al [24] to estimate the false positive rates in Bloom filters (which is computed over all possible inputs to the Bloom filter).…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, we focus on the proposal due to Mullin et al [24] to estimate the false positive rates in Bloom filters (which is computed over all possible inputs to the Bloom filter). More specifically, we compute the false positive rate of a filter B(M, Pt) which has m elements, P f (m), as follows [15]:…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

On the privacy provisions of Bloom filters in lightweight bitcoin clients

Gervais

Čapkun

Karame

et al. 2014

Proceedings of the 30th Annual Computer Security Applications Conference

104

View full text Add to dashboard Cite

Lightweight Bitcoin clients are gaining increasing adoption among Bitcoin users, owing to their reduced resource and bandwidth consumption. These clients support a simplified payment verification (SPV) mode as they are only required to download and verify a part of the block chain-thus supporting the usage of Bitcoin on constrained devices, such as smartphones. SPV clients rely on Bloom filters to receive transactions that are relevant to their local wallet. These filters embed all the Bitcoin addresses used by the SPV clients, and are outsourced to more powerful Bitcoin nodes which then only forward to those clients transactions relevant to their outsourced Bloom filters.In this paper, we explore the privacy of existing SPV clients. We show analytically and empirically that the reliance on Bloom filters within existing SPV clients leaks considerable information about the addresses of Bitcoin users. Our results show that an SPV client who uses a modest number of Bitcoin addresses (e.g., < 20) risks revealing almost all of his addresses. We also show that this information leakage is further exacerbated when users restart their SPV clients and/or when the adversary has access to more than one Bloom filter pertaining to the same SPV client. Motivated by these findings, we propose an efficient countermeasure to enhance the privacy of users which rely on SPV clients; our proposal can be directly integrated within existing SPV client implementations.

show abstract

Section: Introductionmentioning

confidence: 80%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

On the privacy provisions of Bloom filters in lightweight bitcoin clients

Gervais

Čapkun

Karame

et al. 2014

Proceedings of the 30th Annual Computer Security Applications Conference

104

View full text Add to dashboard Cite

show abstract

“…For example, if a 25 kB bit array (200.000 Bits) is already filled with 10.000 elements using 10 hash functions, the false positive rate will be 0.00009. Recent work [4] suggests that this classic formula is wrong for small values of m, but the error can be neglected if m is large enough (> 1000).…”

Section: False Positives and Location Coveragementioning

confidence: 99%

Geocookie: A Space-Efficient Representation of Geographic Location Sets

Ruppel

Küpper

2014

Journal of Information Processing

View full text Add to dashboard Cite

Abstract:A new approach is presented that allows a user agent to compute a compact browser cookie-like character string (called Geocookie) for a third party, whereby for any given geographic location the third party can either infer that the location is definitely not covered by the Geocookie, or it can infer that the location is probably covered by the Geocookie, depending on whether the user agent has or has not included the location beforehand. The approach extends the concept of a Bloom filter and combines it with Geohashes, thus making it possible to store information about visited geographic locations in the filter. Geocookies can be used in many different scenarios for location-based queries and location-based services, whenever a user agent wants to inform a third party about a set of visited locations such that the third party can compile a result that either favors or excludes these locations. In contrast to existing approaches such as session cookies that are mapped to server-sided stored location trajectories, Geocookies provide a compact and privacy-preserving structure which does not reveal the actual set of all visited locations, but provides a one-way check function which can be used by the third party to evaluate given locations against the Geocookie. In addition, Geocookies provide plausible deniability in case of location matches. This paper introduces a formal definition for Geocookies together with a discussion on practical applications and embedding into HTTP headers.

show abstract

A Bloom Filter Based Model for Decentralized Authorization

Foley

Navarro‐Arribas

2013

Int. J. Intell. Syst.

View full text Add to dashboard Cite

A decentralized authorization mechanism is proposed that uses Bloom filters to implement authorization delegation. This lightweight mechanism is unlike conventional approaches that typically rely on public key certificates to implement distributed delegation. In taking an approach based on one-way hash functions, the mechanism may be preferable for use in computationally constrained environments where public-key cryptography is not desirable. C 2013 Wiley Periodicals, Inc.

show abstract

A new analysis of the false positive rate of a Bloom filter

Cited by 95 publications

References 7 publications

On the privacy provisions of Bloom filters in lightweight bitcoin clients

On the privacy provisions of Bloom filters in lightweight bitcoin clients

Geocookie: A Space-Efficient Representation of Geographic Location Sets

A Bloom Filter Based Model for Decentralized Authorization

Contact Info

Product

Resources

About