Google Inc.

Sangho Yoon

Wangyuan Zhang

Pramod Gaud

Jay Yagnik

Vivek Kwatra

Mei Han

Sandeep Singhal

Arash Baratloo

Guangda Lai

Zhifan Zhang

Gael Hatchue Njouyep

Xudong Shi

Vivienne Zhang

Kaifeng Yao

A method includes receiving data objects, determining a predicted lifespan of each data object, and instantiating multiple shard files. Each shard file has an associated predicted lifespan range. The method also includes writing each data object into a corresponding shard file having the associated predicted lifespan range that includes the predicted lifespan of the respective data object and storing the shard files in a distributed system. The method also includes determining whether any stored shard files satisfy a compaction criteria based on a number of deleted data objects in each corresponding stored shard file. For each stored shard file satisfying the compaction criteria, the method also includes compacting the stored shard file by rewriting the remaining data objects of the stored shard file into a new shard file.

Packing objects by predicted lifespans in cloud storage

A system and method for storing data in a distributed network having a plurality of datacenters distributed over a plurality of geographic regions. The method may involve receiving data, including metadata, uploaded to a first datacenter of the distributed network, receiving access information about previous data that was previously stored in the plurality of datacenters of the distributed network, predicting one or more of the plurality of geographic regions from which the uploaded data will be accessed based on the metadata and the access information, and instructing the uploaded data to be transferred from the first datacenter to one or more second datacenters located at each of the one or more predicted geographic regions.

Access pattern driven data placement in cloud storage

Clustering algorithms such as k-means clustering algorithm are used in applications that process entities with spatial and/or temporal characteristics, for example, media objects representing audio, video, or graphical data. Feature vectors representing characteristics of the entities are partitioned using clustering methods that produce results sensitive to an initial set of cluster seeds. The set of initial cluster seeds is generated using principal component analysis of either the complete feature vector set or a subset thereof. The feature vector set is divided into a desired number of initial clusters and a seed determined from each initial cluster.

Principal component analysis based seed generation for clustering analysis

Growing community of inventors

Palo Alto, CA, United States of America

Sangho Yoon