UPDATED 11:29 EDT / FEBRUARY 19 2013

As Facebook Becomes a Smarter Social Service, Hybrid Storage is Key

Storage is strategic IT decisions that companies must make. This is even truer with the rise of the cloud. What data can be stored online, and how does one go about optimizing the available space?

The cloud in its current state is more an extension of functions with advanced backup disaster recovery, although in the near future, we would actually further our goals by storing some data in the primary cloud. The standing trend is to automatically prioritize storage. Thin provisioning technologies, or multi-tiering, prioritizes data based on its importance.

As the enormous volume of data continues to grow at an exponential rate, companies like Facebook plans to build data centers with flexibility in mind, hosting different types of storage, server hardware and networking equipment that use less energy and cost less than the current data center, and without changing the response times.

More content = more storage: movies, books & apps

.

Facebook infrastructure now spans four data centers, tens of thousands of computer servers and more software tools to accompany the growing demand behind the social network.

When the company first launched Facebook’s open ecosystem of applications, games were the first category, which quickly gained momentum in the social network thanks to developers like Zynga. Gaming is still multi-billion dollar business for Facebook, but the company sees the prospects for growth in the market of literature, cinema and fitness tools.

Expanding beyond Facebook’s internal network, the subsequent wave of content-related demand came from news and media organizations, thanks to the “Like” button.  According to Facebook’s VP of Partnerships Dan Rose at the D: Dive Into Media conference, users immediately began to share tons of news and articles in their feeds. Then, with the development of applications such as Spotify, music, too, became a social experience on a large scale.

But now that these categories have won us over, according to Rose, the next step in the development of applications is likely to incorporate movies, books and fitness applications.

What’s more content mean for Facebook’s data center?

What’s more content mean for Facebook’s data center?  Product recommendations based on the efficient monitoring and analysis of socially enhanced data.  Similar to the deal between Facebook and Spotify, Netflix is gearing up to share content on Facebook. Rose said the company wants to allow partners to leverage the data and the preferred content that users share, and allow developers to improve recommendations.

As for books, the social network is interested in applications such as GoodReads, which also allows you to share information with friends about interesting new books.

But with the rise of integrated applications, the increased content has driven demand for Facebook-derived services.  The data center is key towards Facebook’s ability to run Big Data analytics, and perhaps the most revolutionary product to emerge from the network’s capable infrastructure is Graph Search.

In early January, the company announced the upgrade of its search service, Graph Search, which allows users to find information about people, places, photos and interests.  Facebook’s offered little insight to the build out of its Big Data needs regarding Graph Search, but we do know it’s built on a platform Facebook internally refers to as Unicorn.  We also know Facebook’s a big supporter of Hadoop, which works best in uniquely designed data centers such as Facebook’s, built for a future of adaptability.

Hybrid storage paves the way for Facebook’s Services success

.

In the three pillars of the data center – applications, networking and storage – the focus tends to be on the first two, and how changing architectures amongst a variety of factors is helping drive the pace of adoption. But the rise of mobile devices, apps and services has forced companies to look at infrastructure change, with a focus on storage. The very architecture of storage is an all-important aspect of social media business models.

Companies like Google and Facebook are looking for ways to improve the imbalance between system processing needs and storage system capabilities, while maintaining a balance between optimizing two primary storage metrics. And the answer lays in hybrid storage solutions.

Hybrid storage architecture allows these companies to simultaneously use the speed of SSDs in conjunction with SATA disk capacity to offer large storage capacities of data access speeds hundreds of times above the market norm. It can provide near unlimited scalability and is simpler, easier and less expensive to implement.

In light Big Data trends, Facebook’s robust analytics offerings should be designed to bring together data from the numerous platforms mentioned above. To meet this challenge, Facebook can bundle a hybrid storage solution together with software-as-a-service apps, custom-tailored to mesh with its on-premises storage.

Social rewards to hybrid storage

The advantages are manifold: Hybrid Storage is designed to deliver efficient $/GB and $/IOP; SSD and HDD resources virtualized providing performance as a utility to applications; multi-protocol support for iSCSI and Fibre channel coupled with merging NAS and SAN storage into a single platform enabling efficient storage consolidation; reliability and error reduction; scalability; cost reduction; quick technology refreshes and data migration and deployments and on-premises storage can interact with the cloud app as if it was native cloud storage. Wikibon research shows here why a true hybrid storage system is less expensive than their traditional storage.

In addition, there are tremendous performance advantages to on-site analytics and storage. For one thing, the analytics engine and the data are right next to each other, which will be helpful for Facebook to meet its new requirements.

That’s why Facebook is building tools like Scuba and Prism. Scuba platform lets instantly analyze data describing dynamic view of real time data of the company’s massive infrastructure.

Traditional Hadoop solutions help Facebook maintain the world’s largest Hadoop cluster but failed to overcome latency issue. Prism is designed to improve and expand Hadoop and hybrid storage pools can help solve latency problems with targeted solutions.

Clearly 2013 is going to be a year of dynamic changes regarding data center transformations, where meeting the mounting challenges of storage will be vital. The long-term answer seems to be hybrid.


A message from John Furrier, co-founder of SiliconANGLE:

Your vote of support is important to us and it helps us keep the content FREE.

One click below supports our mission to provide free, deep, and relevant content.  

Join our community on YouTube

Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.

“TheCUBE is an important partner to the industry. You guys really are a part of our events and we really appreciate you coming and I know people appreciate the content you create as well” – Andy Jassy

THANK YOU