World’s largest social discovery site uses Object Storage

Massive Media has been utilizing a storage infrastructure built atop Caringo CAStor® for the past year to help quickly and efficiently store and serve millions of photos end users upload to its social networking service Netlog and its social discovery platform Twoo.

Netlog is an online community of more than 110 million members that allows users to create their own web page with a blog, pictures, videos, events, playlists and more to share among their friends. Twoo is the largest social discovery site in the world, according to ComScore personals, and attracts nearly 10 million monthly unique visitors interested in utilizing the site’s innovative matchmaking algorithms to help them meet new people. Combined, the two sites have generated over 600 million thumbnails and photographs. With users currently uploading more than a million additional objects a day, the capacity requirements of Massive Media’s network are expected to double every two years.

Massive Media initially launched with a self-designed storage system then switched to MogileFS, neither of which provided the scalability the company needed. A move to a solution by Amplidata also didn’t work well as the sites grew. That is when Massive Media turned to Caringo to provide the capacity, performance and scalability that it needed.

“At first, everything was going OK with our previous storage, but as we grew, it started to have its hiccups,” said Nicolas Van Eenaeme, Director of Technology for Massive Media. “Also, there was a huge storage overhead in the previous systems we used and all of them had in one way or another a single point of failure. The software design of CAStor is really superb. The shared-nothing architecture is great. The performance is also outstanding. And it scales very well. If you need more throughput or storage, you just add more nodes. It’s clear that CAStor is a winner.”

Massive Media stores each object multiple times on three clusters: Cluster 1 consists of 60 nodes on 30 physical servers, and 7,200RPM SATA drives; cluster 2 consists of 32 nodes with solid state drives; and cluster 3 consists of 42 nodes, with 7,200RPM drives. Nodes in each cluster are interconnected with 1GbE connections. The company needed a solution that could match the throughput and performance requirements of hundreds of concurrent reads, writes and deletes per second. Software updates had to be done with no overall system downtime. Recovery time of a single node had to be less than a day. And, it needed to provide linear scalability in terms of storage capacity with near-linear scale in terms of throughput capacity.

Van Eenaeme said he appreciated CAStor’s ease of setup and integration, which took only hours before everything was up and running. He also liked Caringo’s flexible, ‘pay-as-you-grow’ licensing model, which eliminated the requirement for huge upfront investment costs and helped satisfy the company’s need to provide continued growth while containing costs. Massive Media is now migrating all of its legacy storage to Caringo.

“The explosion in content created by social media requires a solution that provides organizations like Massive Media the means to cost effectively manage millions of objects per day, all of variable sizes,” said Mark Goros, CEO of Caringo. “Our object storage software, powered by CAStor, makes it easy to manage the massive scalability requirements of social media to meet even the most-demanding storage challenges. By offering plug-and-play scalability for both capacity and performance, CAStor allows Massive Media to stop worrying about managing its infrastructure and allows it to focus on providing the best photo-sharing capabilities for its websites.”