Data Monetization is leveraging an organization’s existing data to either increase profits or efficiency. This data can come from a variety of sources including modern devices like the Internet of Things (IoT) and sensors and cameras. But it can also come from legacy sources like users and databases. There is also a wide selection of modern platforms that process and correlate this data such as Cassandra, MongoDB, HBase and Hadoop. Monetizing data sounds like a great idea, who wouldn’t want to make money from data they are storing? The problems are getting the data from the source to the platform that will process it, the cost of storing the data long-term and making sure that the data is readable when it is needed.
In a perfect world you want to design a single storage infrastructure to collect the data from various sources, as well as deliver the data to platforms that will process it. Enterprises considering ways to monetize data have to face two choices. Leveraging the existing NAS architecture or the storage capabilities of the development platform. Both of these choices fall short when addressing the enterprise’s needs and force IT to develop workarounds that increase the cost of the project and decrease its deficiency. Object storage may be the solution to the problem. These systems allow IT to meet the data monetization request without breaking the budget.
Problem 1: Data Collection
The problem with data collection is the devices acting as the source don’t agree on what the target storage system should look like. Most expect an NFS mount point but a few expect an SMB mount point and several now can write to an object storage system directly. For example, a photo sharing site would need to support data coming in directly from a smartphone, from a laptop or from another cloud service. The storage system for a data monetization project needs to be able to not only accept data from a variety of sources but also present data to a variety of requesting applications.
Problem 2: Cost Effective Data Storage
Monetizing data typically requires long term retention of that data. Essentially the more data that the organization has the more likely it can use the data to create a new product or increase efficiency. Imagine the photo sharing site offering to create a birthday card with pictures from the last 20 birthdays that the user had previously uploaded or that a friend of the intended recipient had uploaded and tagged. Without that data the site would have to count on the user to find the uploaded pictures. Now the site can automatically present the data and even create a mock-up of what the card would look like.
The challenge in this example is now storing those pictures in their original resolution for twenty years, multiply that requirement across millions of users and suddenly there is a massive storage requirement. In addition there is no guarantee the user will buy the product, so there needs to be a balance between long term storage and cost effectiveness. The storage system for a data monetization project not only has to scale infinitely, but do so cost effectively.
Problem 3: Data Verification
Finally, making sure the data is still readable after years of storage is critical. Sticking with the photo sharing site example, if two or three years worth of birthday pictures becomes unreadable for some reason, it diminishes the value of the card. The storage system needs to be able to verify data on a continuous basis and self-correct degrading data.
There are storage solutions for companies looking to monetize data. As we discuss in our webinar “Designing Storage and Apps to Enable Data Monetization” object storage systems make a strong case to be that storage system. Watch our on demand webinar to learn the use cases for data monetization, the problems legacy storage systems face hosting a data monetization project and some solutions that will overcome these challenges.