Object storage systems ease data capacity and archival concerns

Storage in the cloud

The public cloud was the first object storage use case to put the technology in the spotlight. Bringing the public cloud economic model on premises has become the Holy Grail for many companies. With a pay-as-you-go model that CFOs crave, the public cloud model has reignited the build vs. buy IT debate. However, cloud goes beyond just economics.

Many legacy storage systems fail the test when it comes to three key metrics: economics, ease of use and scalability. If your storage needs set a high bar for these metrics, you might want to consider an object-level storage system, which can bring a public cloud-like experience to your private data center.

On the flipside, you must be running workloads that are suitable for object storage. For example, while there are object storage systems that can support virtual machines, you might be better off with a block- or file-based system.

Deploying sync-and-share capabilities

Everyone loves Dropbox ... except for IT security people. This kind of sync-and-share service has become increasingly popular as organizations have become more distributed, and as people have acquired more devices. It is not unusual for someone to have a desktop, a laptop, a phone and a tablet, and to need to access work files from all four devices.

That said, there is a desire to maintain strict controls around such file sharing, particularly in industries storing sensitive data. That leaves many companies looking for options that can grow with the company while providing the kind of accessibility enjoyed by Dropbox users.

Object storage vendors are building such capabilities atop their platforms, making it possible for end-user organizations to deploy sync-and-share services while maintaining centralized control and logging. If you need such services, an object storage system may be in your future.

Managing tons of unstructured data

In 2015, retail analytics company Manthan released an infographic that claimed that "every 2 minutes, we generate an equal amount of data to what was created from the start of time to the year 2000." Two years later, it is likely that this pace of data creation has increased.

No matter how you put it, we are creating data at a furious pace, and we need a place to store it. You may have all kinds of data needs in your organization that could be well-served through the adoption of a scalable object storage system.

Images. Suppose you are an insurance company, and you need to store several years' worth of images of damaged homes or vehicles. There are a couple of challenges here. You need to be able to easily grow the storage capacity footprint, and these photos must be easily searchable. Object storage systems offer these capabilities through easy scaling and rich metadata handling features.

Videos. Law enforcement officers wear body cameras. Schools have video cameras in every nook and cranny. The streets are lined with traffic cameras. All of this footage needs to be stored and, like images, videos need to be tagged with identifying information to make them easily searchable in the future. They also need lots of capacity. All of these factors make video surveillance a prime object storage use case.

Big data and internet of things data. Data is the lifeblood of the business. In a world in which the volume and velocity of data are growing at sometimes alarming rates, storage is critical.

Today, organizations have moved far beyond simply gathering customer information; they actively capture constant streams of data from all over the world as their customers install more and more devices in their homes. As is the case with all other kinds of data, this sensor data needs a place to live, and the scalability and ease of use of object storage make it an attractive option for these purposes.

Addressing the need for API-accessible storage

Automation is gaining traction across industries as companies seek to increase output without having to spend more money to do it. IT practitioners and decision-makers are looking for ways to automate IT and other business processes. Traditionally, these kinds of business process re-engineering efforts have focused on the database side of the equation.

However, as processes have become more intertwined with different kinds of data, simple database-centric efforts may no longer be sufficient. You may need the ability to deeply integrate your automation platform with your storage system so that you can directly access files, such as images and video files, where it makes sense, and without the performance overhead of a full database.

In general, object storage systems feature comprehensive APIs. These APIs make storage available as just another programmable object. In the DevOps world, this idea of infrastructure as code is a pervasive and important one.

Consider the following scenario: You have a custom sales platform and store customer records as objects in your storage system. Around the holidays, and as sales increase, you begin to run out of storage. To address this, you can modify your sales platform to watch over storage and move stale customer records to secondary storage or to a cloud archive, keeping primary storage available for paying customers. Once you build this API integration the first time, you will have automated this process, and will not have to worry about it again.

Maintaining availability with always-on storage

Thanks to replication and erasure coding, object storage can maintain high levels of availability. With replication, the storage system makes automated copies of data, either locally or in other data centers. Different object storage systems have different capabilities. Some can make two or more copies, and some are configurable based on your needs.

Erasure coding is a complex data protection technique that involves breaking data into a series of chunks. Your system might break data into 12 chunks, augmented by four parity blocks, for a total of 16 blocks, which are then distributed to various locations. In the event of a hardware failure, you could recover your data by reading any 12 of those 16 blocks.

Object storage is great for relatively static data that does not change a lot, and that is not accessed randomly with an expectation of massively high performance. This makes using archival applications another obvious object storage use case. Archived data stays mostly consistent and does not need to be constantly read back.

Where object storage fails, in many cases, is when it comes to supporting virtual machines and relational databases. Both of these workloads require a lot of data change and a lot of random data access. You never know what data you will need next. So, in short, if you need relatively static archival storage and do not plan to run virtual machines or relational databases on it, object storage may be a good option for you.