Sketch-based change detection: methods, evaluation, and applications

By Balachander Krishnamurthy, Subhabrata Sen, Yin Zhang, Yan Chen. ACM/SIGCOMM Internet Measurement Conference, 2003

663 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/35a25052bf9fc89e0b28499236322c5786fc58a9

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Statistical and Nonlinear Physics

Citations in the corpus

Change-point monitoring for the detection of DoS attacks (2004, neutral): “, wavelet-based [1], spectrum-based [21], sketch-based [29], and signal processing approaches [57].”
Diagnosing network-wide traffic anomalies (2004, methodology): “The same general strategy of seeking a useful alternate basis set for data representation lies behind previous approaches to anomaly detection as well [1, 2, 14].”
Structural analysis of network traffic flows (2004, methodology): “A number of change detection methods have been proposed that rely on wavelet denoising techniques [1] and deviations from forecasted behavior [3, 14] to identify outliers.”
Data streaming algorithms for efficient and accurate estimation of flow size distribution (2004, methodology): “For example, they have been the building blocks for detecting traffic changes [19], identifying large flows [5], and constructing Bloom filters [20] that allow both insertions and deletions (called counting Bloom filter) [21].”
Distributed data stream processing and edge computing: A survey on resource elasticity and future directions (2017, neutral): “do not attempt to model application behaviour or detect changes in the incoming workload (Krishnamurthy et al., 2003) – and offer methods”
Mining anomalies using traffic feature distributions (2005, positive): “Other work proposes sketch-based methods to detect traffic volume changes and hierarchical heavy-hitters [22, 35].”
One Sketch to Rule Them All: Rethinking Network Flow Monitoring with UnivMon (2016, methodology): “Here, custom online algorithms and data structures are designed for speciﬁc metrics of interest that can yield provable resource-accuracy trade-offs (e.g., [17,18,20,31,36,38,43]).”
Open challenges for data stream mining research (2014, background): “The short length of tweets in particular creates added complexity to text-based search tasks as it is difficult to identify relevant tweets matching a user query [13,35].”
Unsupervised Anomaly Detection via Variational Auto-Encoder for Seasonal KPIs in Web Applications (2018, methodology): “, [6, 17, 18, 22, 24, 25, 29, 37, 38]), with in total 133 enumerated configurations of hyper-parameters for these detectors.”