Dispel decrypt distributed cloud storage tool Fastdfs

2012 The database technology conference was held in Beijing on April 13 at the Sheraton Hotel. The Conference has been the third database technical conference sponsored by IT168, Itpub, Ixpub and Chinaunix. The conference will focus on the large data architecture design, database security, distributed database, business Intelligence, NoSQL, Hadoop and many other key topics. The Conference has been the national database technology experts of high concern and support, is the current symbol of the highest level of the database engineer event.

In the afternoon of 14th distributed large data Architecture Design special speaker is open source Fastdfs developer Qing. Qing, with over 10 years of internet development and architecture experience, has been in the Sina network as a development engineer, in Yahoo China as an architect, now a Amoy network engaged in data capture related work. Qing has been in charge of the design and development of Sina Content Management system, Yahoo Forum and Yahoo album, and has very rich experience on large web site architecture and storage. Qing is the author of open source Distributed File System Fastdfs and distributed KV system FASTDHT, and has a deep research on distributed data storage architecture.

Qing

Qing first to introduce Fastdfs, he said Fastdfs is an open source lightweight Distributed File system, it manages the file, features include: file storage, file synchronization, file access (file upload, file download), etc., to solve the problem of mass storage and load balancing. Especially suitable for documents as the carrier of online services, such as photo albums, video sites and so on. The FASTDFS server has two roles: Tracker (tracker) and storage node (storage). The tracker mainly does the dispatching work and plays the role of load balancing on the access.

Features provided by Fastdfs

Characteristics of Fastdfs

Fastdfs Extended Module Features

Mr. Qing also told us how Fastdfs solves the problem of synchronization latency. He says Fastdfs in the file name of the stored build contains the source storage IP address and the file creation timestamp. At the same time source storage timing to tracker report synchronization, including to the target server synchronized to the file timestamp, and tracker received storage synchronization report, find the group of each storage synchronized to the time stamp (take the minimum), Saved in memory as the storage property. This minimizes synchronization latency.

Finally, Mr. Qing introduced the FASTDFS cloud storage deployment scheme, the details of the following figure.