Object Storage In The Public Cloud

How are cloud technologies such as object storage transforming business?

Enterprise IT and service providers are transforming how they do business by leveraging cloud architectures to deliver infrastructure-as-a-service (IaaS). Cloud architectures are ideally suited to the unique storage demands of unstructured data, which is forecasted to grow at 80%. This growth is driven by the increasing size and amount of digital images, video, audio, and a combination of these media as seen on sites such as Amazon and others.  Consumers routinely use the richness of these differing media streams to get information and use the written word for only a small fraction of their information intake. However, until recently, business users have been limited to perusing documents alone.  Now, technology has finally caught up with how we digest data; in particular, object storage is the fundamental transformative technology underlying cloud architectures and enabling the business transition to IaaS.

The modern world is fast paced and instant gratification the norm. Increasingly, cloud technologies like object storage fulfill the gaps left by block-based and file-based storage. These technologies deliver cost savings and drive new revenue streams.

Information On-Demand

If you’re over 40 years of age, you probably recall a time when you had to wait to watch a television show. If you missed a broadcast, you might get to view the missed show as a summer rerun.

Fast-forward to today and it is an on-demand world. DVRs and DVD/Blu-ray players have transformed TV viewing, enabling you to watch what you want, when you want. Streaming is the next phase as TV programs bypass hard media devices such as DVR players, and go straight from the Internet to your viewing medium of choice via online sites like Netflix, Amazon, Hulu, and the television networks themselves.

Interestingly, the online sites do something different than the networks that both produce and also stream the programs. The online sites augment their offerings with predictive analysis engines that enable them to “push” additional programming to you. This is done by applying intelligence to viewing patterns to suggest new titles, rather than just relying on the “pull” of users deciding on their own to view certain programs.

Keeping It in Context

Object storage systems support per-object metadata or data about data. Meta data gives context to stored data in relation to other data and establishes business value for the objects. This approach is not as easily implemented in block-based and file-based storage systems due to their hierarchical rather than associative data structures.

For example, TV viewership numbers on their own have limited meaning but have considerable value when associated with a specific customer. In the TV-on-demand scenario above, individual views mean less than what the multiple data points of many viewed titles say about the customer. By analyzing the categorization stored in the meta data of the TV shows a customer has watched, it is possible to determine which types of shows that particular customer prefers. With this knowledge, the video service provider can automatically push new suggested titles to the viewer to grow consumption.

The Internet and Scale-Out Object Architectures

What’s changed to make this possible?

The rise of the Internet and scale-out object architectures has changed how information gets consumed. These changes include:

  • Object IDs: Object storage introduced unique IDs or URLs for each stored aggregate of data, including online documents, images, videos, and more. These unique IDs plus the meta data associated with each object means data can be stored anywhere for retrieval via standard Web-based protocols. The name of each object is universal and unchanging, without any dependence on an individual storage system or naming hierarchy.
  • Open APIs /HTTP protocols: APIs based on SOAP/REST design models and HTTP protocols made seamless, global access possible. Mobile devices, tablets, and wireless networking opened up even more options for receiving the abundant amounts of information available.
  • Lower cost x86-based technology: The new API design models and HTTP also made the use of commodity (x86-based) processors, memory, and networking possible. This change enabled management to move from the devices to software to ease administration and lower the overall total cost of ownership.
  • Abundant bandwidth: The Internet spurred a massive build-out of connectivity and networking infrastructure, making massive aggregate bandwidth available.
  • Active-active configurations: These technologies also enabled active-active configurations where the aggregate bandwidth supports multiple access points; a key requirement in cloud scale deployments where there is always a node or disk down.

Object Storage Transforming Online Commerce

How can object storage be applied to industries?

Object storage use cases can be found in many industries, including healthcare, finance, and other online commerce, in addition to television programming. Some other use cases include:

  • Healthcare:  Medical records and images on their own have limited meaning but have considerable value when associated with a specific patient, medical condition, or demographic.
  • Financial Services: Bank balances, mortgages, and credit ratings alone provide information about a customer, but combined enable targeting home owners and others for new financial services products like home equity and auto loans.
  • Online Shopping: Understanding buying habits can lead to the provider offering additional services like, for example, music recommendations for the audiophile; Amazon emails me with new music selections about once a week based on my buying history with them.

Object Storage Opening Up New Services for Public Clouds

Adding object storage to service-provider public clouds provides an attractive alternative for organizations looking to leverage object storage and data analytics to create new revenue streams.

Time to market for the services provider can be quick too. For example, $100 billion ($USD) telecommunications services provider giant AT&T decided to leverage its extensive communications networking backbone to deliver IaaS to its enterprise customer base. The challenge was to bring object storage to market without incurring the development costs associated with introducing new technologies. The solution was to deploy EMC Atmos cloud storage platform to store, manage, and protect globally distributed, unstructured content at scale in multiple data centers. Out-of-the-box Atmos enabled AT&T to set policies for different service levels and integrate an object-storage services catalog including a self-service user portal, reporting, and metering with AT&T billing systems. The result was a new business model with self-service object storage for enterprises in 90 days.

Object storage offers much to the public cloud, and also enables new possibilities for organizations looking to implement private cloud models. Future posts will address object storage in the enterprise including what to consider when deciding whether or not to outsource object storage or to keep it in-house.

About the Author: Mark Prahl