Over the next few years enterprises will face significant storage challenges as traditional file system storage fails to keep up with the complexity of ever-increasing file counts, file sizes and varying data types.
Organisations are going to extreme lengths to tackle issues related to the retention, management and retrieval of data that go hand in hand with the ubiquitous NAS (network attached storage) design, which has been in place for the past two decades and is showing signs of reaching breaking point.
NAS systems have previously stored home directories and data sets such as project files and user-created files adequately, but the onset of more unstructured data, machine-generated data, a by-product of the Internet of Things, means that reliably retaining and managing new data types cost-effectively is proving to be a real challenge for both SAN (storage area network) and NAS systems.
Today’s models for storing data by directory and filename are ultimately ineffective and non-sustainable in this big data era. So creating a long-term data storage strategy to embrace a new construct capable of future proofing the storage environment is worth exploring.
Object storage is rapidly becoming recognised as a compelling, simple and scalable solution to the storage pains big data growth has caused, facilitating ease of information access, recovery and protection.
Object storage particularly shines when it comes to storing and supporting the explosion of unstructured data such as images, video, music, bio-informatics and a plethora of file types, all on a massive scale.
With its integrated metadata and tagging functionalities, locating data can be accomplished with ease. Since there are two primary reasons to store long term, for legal or financial benefit, the monetisation of data will be the focus for many future-facing organisations, the adoption of object storage technology is one that makes sense.
Object storage adoption
So where does object storage stand in terms of adoption? A selection of content-intensive vertical markets, such as media and entertainment, healthcare, geology, genomics, technology, surveillance and engineering have already started to implement object storage technology. Beyond this, awareness of the benefits and end user expectation is still relatively low, but shows signs of growth.
According to IDC, in 2014, OBS (object-based storage) solutions accounted for ‘nearly 46% share of the file-and-OBS (FOBS) market in revenue and is forecast to be a $28.3 billion market in 2018. However, the total revenue from commercial OBS products as a percentage of the overall OBS market remained quite small in 2014, and even with a forecast CAGR of over 27%, it will constitute a relatively small portion in 2018’.
The benefits of object storage come firstly in its ability to store data and metadata in the same system. This allows for a single name space and system to be searched for relevant data. The platform also enables numerous ways of accessing data.
Another big advantage is scale. Object storage scales to the billions if not trillions of objects, whereas current file systems begin to strain at around 100 million files. Object storage will prove a compelling option for organisations seeking to rein in their storage spend, as storing unstructured data on an object tier will mean significant cost savings when compared to tier 1 SAN or NAS.
Most object storage technology is also highly resilient to drive failure and is known for its rapid recovery capabilities. While best known for having a disk back end, when you dig deeper into the technology you see that it is actually storage device agnostic and that anything from flash, disk, tape or even optical could be used.
Today, tape has the best cost per byte stored ROI. This makes tape the perfect choice for an object storage archive, still having an object storage front-end, meta data stored in flash, with an SLA of time to data that is slightly longer than traditional disk or flash-based object storage.
Along with this, an object storage archive allows for the primary object store to lazily backup to the archive object store thereby deploy a tiered object or second copy.
Barriers to adoption
With so many of today’s systems built around the traditional file system access, it is not surprising that moving to object storage will seem like a daunting task, as it does require a change of mindset.
There is no longer the comfort of a mounted file system as well known as NAS. But this challenge can be looked at in the same way that the internet become integrated into our lives.
Looking back 25 years, one of the very reasons it came into being was to share data. Using the same protocol that the WWW uses now to move data is nothing new, it has just massively evolved from storing open text formatted data on a web-page to complete storage solutions.
So although implementing an object storage system will require some changes to be made in the system architecture, once started, it will allow systems to scale far beyond what is currently being deployed.
As more organisations begin to use object storage the benefits of changing will far outweigh the small change to implement. And just as NAS and a clustered file system can co-habit, object and NAS storage can co-exist in the same environment.
For organisations considering a technology refresh but unsure of whether to make the move to an in-house deployment can be reassured that its benefits can also be gleaned through the services of third-party cloud vendors utilising object storage.
Object storage planning
So how do CIOs and IT directors go about planning for object storage deployment? Since object storage is an evolution, not a revolution, planning needs to start at the top, to make sure it meets the business need.
Followed by a review of how object storage can not only help reach that goal but increase profit by storing data that is more searchable, accessible and therefore easier to monetise.
Object storage allows a much tighter integration and connection to the business logic and the data needed to drive the business.
For successful implementation, the process should include conversations with relevant departments, specifically with the people who create or manufacture the product, which in this case is data, for example, editors or scientists.
Secondly, a discussion should be had with the core services team, the team who delivers IT infrastructure.
CIOs who look ahead, ask the right questions about the business needs and make the necessary provisions for deploying object storage (if it proves to be an appropriate choice for their businesses), will not only create ample storage and management capabilities, but aptly position the organisation to create new revenue streams fuelled by their data’s value.
Sourced from Matt Starr, CTO, Spectra Logic