How many times have I written about this? Plenty, I know. But it’s actually noteworthy that it’s finally happening… more and more object storage vendors are finally broadening the scope of their products by making them multiprotocol and, consequently, much more interesting for enterprise use cases.
More and more object storage vendors are finally broadening the scope of their products by making them multiprotocol and, consequently, much more interesting for enterprise use cases.
Objects are cool but…
We all love objects, the concept at least, and the ability to associate metadata to data is amazing. But, if you live in the real (enterprise) world and rewriting your applications is not easily done, accessing them through APIs is not that great.
Also, finding developers who want to change the way they interact with storage can be a chore. They’ve always been used to dealing with file systems and, like everybody else, they are slow to make changes… It’s also true that younger cloud native developers are quite keen on objects but are still a minority in enterprise organizations.
Finding developers who want to change the way they interact with storage can be a chore.
So,standard interface is also needed, like SMB or NFS, while maintaining the ability to access the same objects/files via APIs… because, sooner or later, you will need both!
…but we want more now
No doubt Sync and Share, as well as many other mechanisms (like remote NAS gateways) are growing in popularity. They are aimed at helping remote and mobile users access data seamlessly from everywhere and their adoption is directly connected to how dispersed the organization is and not to capacity.
And there’s more… tiering is becoming a common way to manage data between on-premises systems and the cloud. In some cases vendors are building quite an interesting ecosystem around their object stores to help the end user build infrastructures that can perform automatic policy-based tiering between primary and secondary storage.
No doubt Sync and Share, as well as many other mechanisms (like remote NAS gateways) are growing in popularity.
The most unified storage ever
Contrary to what we have seen in the past, the association of block+file doesn’t make sense any longer, it’s an obsolete concept of unified. Now we want huge capacities and throughput to sustain unstructured data workloads and the lowest, consistent latency to access our DBs faster. They just don’t belong together.
This is why the concept of unified storage has to move forward and cover a wider spectrum of protocols and devices by targeting only unstructured data workloads. Data and workloads that can span from local to cloud infrastructures, but maintaining similar access methods.
The concept of unified storage has to move forward and cover a wider spectrum of protocols and devices by targeting only unstructured data workloads.
Closing the circle
There are many object storage systems available in the market now, and it’s clear that most of them are finally aiming at something more than storing your ice-cold data. They absolutely cannot manage latency sensitive workloads, but they can be much faster than in the past (or they can be smaller… after all, there aren’t so many “Facebooks” around).
The number of vendors growing their own ecosystem has been increasing for quite some time now. And we’ll be seeing even more native connectors, access protocols and gateways providing 360 degree solutions, bringing object storage economics and usability to the enterprise in the next few months.
I wrote a new white paper for DDN this month covering this exact topic (You can find it here). DDN is one of those vendors building this kind of ecosystem, They’ve always had a very scalable and robust object storage system… but now, thanks to the integration with their primary storage and the addition of native file access, they are undoubtedly broadening the number of enterprise use cases.