Yin Zhang3221 60Th St, Flushing, NY 11377

7293086, Nov 6, 2007

Jul 25, 2003

10/627767

Nicholas G. Duffield - New York NY, US
Albert Gordon Greenberg - Summit NJ, US
John G. Klincewicz - Wayside NJ, US
Matthew Roughan - Morristown NJ, US
Yin Zhang - Lake Hiawatha NJ, US

AT&T Corp. - New York NY

G06F 15/173

709224

A method and apparatus for the estimation of traffic matrices in a network are disclosed. Mechanisms are disclosed for measuring traffic volume from a plurality of ingress points to a plurality of egress points in a large scanl network, such as an IP backbone network. The traffic matrix is advantageously inferred from widely available link load measurements such as SNMP data.

7574506, Aug 11, 2009

Sep 26, 2007

11/904086

Nicholas G. Duffield - New York NY, US
Albert Gordon Greenberg - Summit NY, US
John G. Klincewicz - Wayside NJ, US
Matthew Roughan - Morristown NJ, US
Yin Zhang - Lake Hiawatha NJ, US

AT&T Intellectual Property II, L.P. - Reno NV

G06F 15/173

709224

A method and apparatus for the estimation of traffic matrices in a network are disclosed.

7751325, Jul 6, 2010

Jun 14, 2004

10/867265

Balachander Krishnamurthy - New York NY, US
Subhabrata Sen - Chatham NJ, US
Yin Zhang - Austin TX, US
Yan Chen - Evanston IL, US

AT&T Intellectual Property II, L.P. - Reno NV

G01R 31/08

370233, 370253

A sketch-based change detection technique is introduced for anomaly detection. The technique is capable of detecting significant changes in massive data streams with a large number of network time series. As part of the technique, we designed a variant of the sketch data structure, called k-ary sketch, uses a constant, small amount of memory, and has constant per-record update and reconstruction cost. A variety of time series forecast models are implemented on top of such summaries and detect significant changes by looking for flows with large forecast errors. Heuristics for automatically configuring the forecast model parameters are presented. Real Internet traffic data is used to demonstrate and validate the effectiveness of sketch-based change detection method for utilization as a building block for network anomaly detection and traffic measurement in large computer networks.

8145745, Mar 27, 2012

Dec 28, 2005

11/321218

Zihui Ge - Secaucus NJ, US
Albert Gordon Greenberg - Summit NJ, US
Matthew Roughan - Erindale, AU
Yin Zhang - Austin TX, US

AT&T Intellectual Property II, L.P. - Atlanta GA

G06F 15/173

709224

Method and apparatus for network-level anomaly inference in a network is described. In one example, link load measurements are obtained for multiple time intervals. Routing data for the network is obtained. Link level anomalies are extracted using temporal analysis on the link load measurements over the multiple time intervals. Network-level anomalies are inferred from the link-level anomalies.

8228803, Jul 24, 2012

Jun 28, 2006

11/477437

Zihui Ge - Secaucus NJ, US
Yin Zhang - Austin TX, US

AT&T Intellectual Property II, L.P. - Atlanta GA

H04L 1/00

370238

Method and apparatus for determining at least one critical traffic matrix from a plurality of traffic matrices, where each of the plurality of traffic matrices is organized into at least one of a plurality of clusters, for a network is described. In one embodiment, a merging cost is calculated for each possible pair of clusters within a plurality of clusters. A pair of traffic matrices that is characterized by having the least merging cost is then merged. The calculating and the merging steps are subsequently repeated until a predefined number of clusters remains, wherein the remaining clusters are used to determine at least one critical traffic matrix.

2016035, Dec 1, 2016

Aug 12, 2016

15/235434

- Atlanta GA, US
Zihui GE - Secaucus NJ, US
Ajay MAHIMKAR - Woodbridge NJ, US
Aman SHAIKH - Berkeley Heights NJ, US
Jennifer YATES - Morristown NJ, US
Yin ZHANG - Austin TX, US
Joanne EMMONS - Howell NJ, US

H04L 12/24

A system and method identify a network upgrade from a data set including a plurality of configuration sessions. The system performs the method by receiving a plurality of configuration sessions. Each of the configuration sessions comprises a plurality of configuration commands. The configuration commands are generated by a same user identifier and within a time threshold. The method further includes identifying one of the configuration sessions as a network upgrade session. The identification is based on a rareness of the configuration session or a skewness of the configuration session.

2016011, Apr 28, 2016

Jan 6, 2016

14/988997

- Atlanta GA, US
- Austin TX, US
Tae Won Cho - Jersey City NJ, US
Yin Zhang - Austin TX, US

AT&T Intellectual Property I, L.P. - Atlanta GA
Board of Regents, The University of Texas System - Austin TX

G06N 7/00

Recommendation systems are widely used in Internet applications. In current recommendation systems, users only play a passive role and have limited control over the recommendation generation process. As a result, there is often considerable mismatch between the recommendations made by these systems and the actual user interests, which are fine-grained and constantly evolving. With a user-powered distributed recommendation architecture, individual users can flexibly define fine-grained communities of interest in a declarative fashion and obtain recommendations accurately tailored to their interests by aggregating opinions of users in such communities. By combining a progressive sampling technique with data perturbation methods, the recommendation system is both scalable and privacy-preserving.

2015010, Apr 9, 2015

Dec 11, 2014

14/567113

- Atlanta GA, US
- Austin TX, US
Tae Won Cho - Jersey City NJ, US
Yin Zhang - Austin TX, US

AT&T Intellectual Property I, L.P. - Atlanta GA
Board of Regents, The University of Texas System - Austin TX

G06F 17/30

707767

Recommendation systems are widely used in Internet applications. In current recommendation systems, users only play a passive role and have limited control over the recommendation generation process. As a result, there is often considerable mismatch between the recommendations made by these systems and the actual user interests, which are fine-grained and constantly evolving. With a user-powered distributed recommendation architecture, individual users can flexibly define fine-grained communities of interest in a declarative fashion and obtain recommendations accurately tailored to their interests by aggregating opinions of users in such communities. By combining a progressive sampling technique with data perturbation methods, the recommendation system is both scalable and privacy-preserving.