In the past, if you needed more storage, you simply ordered it. After all, storage was cheap. Now in the days of cloud computing, you simply provision for more.
However, taking a commodity approach to storage doesn't address the key storage issues that come with big data. With unstructured big data now comprising 80% of corporate data, and with data overall continuing to double worldwide every two years, how do you store and manage all of this data? See below for five ideas.
1. Plan what goes where
Not all unstructured data is used every day. Other data is accessed often and continuously. When large files of unstructured data such as videos and images are in demand, more processing and throughput is required to get them to users. Consequently, storage professionals must decide how to tier this data. Does it get placed in rapid access, solid state memory to afford users immediate access or in mid-level media such as hard drives that provide a medium level of data retrieval? Maybe it should be stored on tape or very slow hard drives that are known as cold storage and are only used for data that is sparingly accessed, or not at all? Storage professionals, with the help of storage automation technologies, must consider and make these type of decisions.
2. Decide what stays and what goes
It's important to ask: What big data stays on premises and what goes to the cloud? And if you need to aggregate this data for purposes of analytics, how do you coordinate secure data integration and transfers between all of these data sources?
Storage professionals need to get involved in several ways. First, they need to lay out a storage architecture that is both on premises and in the cloud. In doing so, they must take into account frequency of data access; and whether accessing the cloud introduces too much latency, making high-demand data transfers impractical.
3. Choose your data
Initially, companies kept every big data payload, because they were concerned about regulatory pressures or a requirement to do a future e-discovery. Individuals in legal, security, or in other user departments normally make these decisions, but it is still incumbent on storage managers to orchestrate a storage architecture to fulfill these requirements. Along with this, decisions need to be made on which data to discard, and when.
4. Plan for disaster recovery
Disaster recovery isn't what it used to be. Today, you not only worry about the systems and data that you manage in your data center—but you worry about what goes on in the cloud, too. If you're using a SaaS (software as a service) vendor, for example, what if the vendor subcontracts to another vendor for its cloud services? If something happens to the data you store at the third party's data center, what do you do since you won't have much leverage? There is also the issue of big data that is collected at the edges of your enterprise. If you're managing storage at remote sites as well as in your data center put security, backup, and failover procedures in place. View More
