Dropbox, Inc.

Microsoft Technology Licensing, LLC

Aakash Kambuj

Vinod Valloppillil

Calvin Low

David W Zhang

Rodrigo Gomez Avila

William R Fraser

Yufei Guo

Bruce Wayne Copeland

Kevin Andrew Chan

Anna Geiduschek

Raveesh Nayar

Dong-Jae Chung

Sam Stafford

Disclosed are systems and methods of accessing content items stored at a content management system by a mobile device, when the mobile device is not communicatively coupled to the content management system. The content management system can receive an indication to mark a content item (e.g., folder and the content items associated with the folder) as available to a mobile device when the mobile device is offline. The content management system can mark the content item, and any content items associated with the content item, for offline access by the mobile device and determine a synchronization schedule for the content item(s). The content management system can then initially send to the mobile device, the marked content item(s) in accordance with the synchronization schedule. While offline, the mobile device can make changes to the marked content items. The changes can be synchronized between the mobile device and content management system when the mobile device is online.

Accessing network based content items by a mobile device while offline

Aspects of the subject technology relate to ways to avoid file duplication in a cloud storage service. In some aspects, a method of the technology includes steps for retrieving, via a computer network, one or more files from a first memory location, identifying a first signature and a second signature for each of the files, and storing the first signature and the second signature into a duplication reference list. In some aspects, the method further includes steps for analyzing files in a second memory location to identify duplicate files using the duplication reference list. Systems and computer-readable media are also provided.

File deduplication using signature information

One or more techniques and/or systems are disclosed for efficiently storing metadata of partitioned data (e.g., directed to a specific individual/user). A first subset of data is identified from the set of partitioned data. The first subset of data comprises data that has a high frequency of access, and may be identified, at least in part, by frequency of access metadata that is associated with the data in the set. Metadata, such as discovery information, permissions, version info, etc., that is associated with the data in the first subset can be stored in low latency (e.g., fast) storage, proving more efficient access. Metadata associated with data from the set of partitioned data that is not in the first subset (e.g., that is not accessed frequently) can be stored in higher latency (e.g., slower, less expensive) storage.

Efficient metadata storage

Growing community of inventors

Seattle, WA, United States of America

Aakash Kambuj