Object Storage is attracting more vendors as the technology moves ‘out of the clouds’ and closer to what could be called “mainstream” IT, at least that seems to be the objective based on the discussions at this year’s NextGen Object Storage Summit (NGOSS) in Los Angeles. According to one of the presentations, all of the major players in storage except IBM have come out with an object storage product (unless you count SmartCloud) either one they developed internally or have licensed from an OEM.
As I covered in my recent column “Object Storage 101”, this is a technology with the ability to give storage systems some compelling advantages over RAID-based disk arrays, particularly those using a traditional ‘scale-up’ architecture. Given the explosion of file-based data, especially files that are created and seldom modified (known as digital “content”) object storage clearly fills a growing need. This is becoming even more true when you consider that much of this content needs to be kept for long periods of time, but still remain accessible.
More specifically, an object-based architecture can enable a file storage infrastructure to expand without traditional file system limits and do so economically, while maintaining relatively fast file access as it grows into the PB range. It also supports long-term data integrity, advanced search, security and other features, plus direct access by applications and the internet.
Object storage is able to deliver these benefits through an impressive list of features and functionality that’s intrinsic to the technology itself. For example, the fact that it encapsulates data in buckets (the objects) which are individually addressable, makes access fast at almost any scale. This same characteristic supports scale-out architectures which are ideal for commodity hardware, keeping costs down. This architecture is also ideal for erasure coding which can replace RAID and support geographic distribution, improving data integrity and reducing the number of copies created for DR purposes. Finally, its ability to store rich sets of metadata within each object brings another set of potential benefits that improve security, access and data relevance.
Mixing “how” and “what”
The way it delivers these benefits (the “how”) is important, but not necessarily in the same discussion as what those benefits are (the “what”). Unfortunately, the confusion factor for object storage in the market has been very high due in part to the fact that the vendors have been mixing these messages. There are products being sold by the vendors at this conference that all use object-based architectures but could be described very differently.
Fortunately, the companies at this conference are aware of that problem and seem to be taking steps to address it. Those companies and their products are (definitions are theirs):
- Amplidata – Amplistore object-based storage platform
- Caringo – object storage software
- CTERA – Cloud storage services platform
- DDN – WOS hyper scale object storage platform
- Exablox – OneBlox object-based storage appliance
- HDS – HCP Content Platform
- Intel – supports object storage industry, testing and standards
- InkTank – Ceph open source object and block storage platform
- NEC – HydraStor scale-out grid storage platform
- Nexenta – NexentaStor software-defined storage platform
- Quantum – Lattus object storage
- Scality – RING object storage software solution
Storage Swiss Take
While there’s certainly some work that needs to be done in the marketing of object storage, the technology is sound and the benefits of the products it can support are impressive. This is evidenced by the diversity of vendors at the NGOSS, from hyper-scale cloud storage to software-only solutions to scale-out NAS systems for small to mid-sized businesses and open source software for anyone who wants to build their own infrastructure. For end users there are and should continue to be solutions available to handle the big data problems (including Big Data) that companies in the clouds and here on the ground will be facing tomorrow, next year and far beyond that.