Andres Rodriguez is founder and CTO of Nasuni and former CTO of the New York Times.
For decades the company data center, housed on site at headquarters, served as the sacred ground in which an organization’s IP was confined, protected and managed. For a time, everything worked fine. For those with offices nearby, flexible latency requirements and limited data volumes, performance was more than acceptable. But it didn’t take long for large, global distributed organizations to find the limits.
Almost from the start, corporate IT departments at multi-national organizations struggled to deliver, control and manage a global file share capable of enabling employees around the world to access, use and store the information they needed. And with the addition of every new office or branch, the complexity of the network increased.
IT found itself bogged down with more servers to maintain, more data to move and more demands to meet. Users of course wanted local performance and immediate access to files regardless of where they resided, but as the edge of the network creeped ever further away from headquarters, everything from file access to disaster recovery grew more difficult. Real collaboration was virtually impossible.
WAN acceleration schemes and other stop-gap measures provided some relief, but they were costly and added even more complexity. Finally, two developments pushed global organizations to the brink. First, data volumes increased exponentially. The three-year hardware refresh cycle that made traditional storage hardware vendors a fortune and forced IT to estimate its future data capacity needs – and buy it in advance – became unsustainable.
Simultaneously, large cloud service providers like Amazon Web Services, Google and Microsoft’s Azure, proved that they could deliver storage capacity at a fraction of the cost. For the first time, the cloud was invited to the party, albeit reluctantly by enterprise IT departments who still considered it risky to let data off premises.
Now things are very different. There isn’t a CIO alive who doesn’t feel the cloud is here to stay. And at the same time, another transformative change is taking shape. Multi-national conglomerates no longer have a monopoly on the need for an IT infrastructure that scales globally. Today, organizations of all kinds have employees scattered around the world that must not only share important data, but also work off of the same files with no fear of overwriting someone else’s work.
So how do you scale your infrastructure for global expansion and what does it get you? For an answer, we must first look at the cloud.
Cloud Object Storage – The New Disk
We are in the midst of the single most disruptive event to ever impact IT. Large organizations are either already in the cloud or are pushing data from hardware-based systems to software-defined alternatives.
It makes sense. Hardware was always the limiting factor, and in no arena was that more evident than in the storage and use of unstructured data. Unstructured data and application files – from the usual suspects like documents and spreadsheets to software code, design files and medical images – are the very lifeblood of the modern enterprise. They must be stored and protected, but as data volumes continued to increase, this became prohibitively expensive and more difficult.
Cloud object storage addressed this issue of capacity. Freed from hardware constraints, IT could spin up or spin down capacity in minutes and only pay for what it used. But object storage did nothing to address the larger questions facing enterprises that wanted to scale their infrastructure globally. It didn’t enable employees to access these files from anywhere as if they were on their own desktop, allow them to collaborate with a colleague a continent away or make it possible for an analyst to do their data crunching. Perhaps most importantly, cloud object storage by itself did little to address even the most basic requirements that enterprise IT departments have for the primary storage of their data.
Enterprise IT Requirements – The Same Vigilance is Required in the Cloud as On-Premises
The leaders of enterprise IT departments are rightfully paranoid. Five years ago, it was difficult to convince many that it made sense to use cloud infrastructure. Now most have come to the realization that in addition to offering unlimited capacity, cloud object storage is also inherently far more redundant and resilient than anything they can put in place on their own.
Of course, the primary storage of files requires far more than capacity and resiliency. To achieve real collaboration and enable global expansion, object storage must also deliver a level of control and performance that equals or exceeds what traditional network attached storage (NAS) systems can deliver in the best of circumstances.
It is in the delivery of these capabilities that a cloud-native global file system comes in. It makes cloud object storage usable, enabling enterprises to utilize private or public cloud object stores for primary storage and more.
For a file system to be global, it needs to not only include the affordable, limitless capacity multi-national enterprises require to keep all of the files, directories and metadata they generate, but also all of the controls and attributes IT departments have to come to expect in years of relying on traditional NAS. These include capabilities a cloud-native global file system delivers even more effectively:
- Military-grade encryption: In addition to being fully encrypted at rest and in transit, only the enterprise should posses the encryption keys for files and other unstructured data kept in the cloud.
- Centralized Control: IT should always have complete control over file access and permissions down to the most granular level, no matter where the files or users are located. With a cloud-native global file system in place, IT can not only determine who needs access to different file workloads, but where they need it and at what level of performance.
- The ability to see and manage all files: IT should be able manage their entire storage infrastructure, including files that are on premises or in the cloud, from a single pane of glass.
- Ironclad backup and recovery: This not only includes setting the recovery point objective for various workloads, but also how frequently snapshots occur. Notably, when coupled with the unlimited, affordable capacity that object storage delivers, IT can take snapshots of the global file share at virtually any frequency desired – eliminating the need for costly traditional backup and disaster recovery infrastructure.
- High performance file access: With a cloud-native global file system, active data is cached on site in a dedicated appliance or virtual machine. Users enjoy immediate access to the files they need as if they resided on their own desktop, while an immutable gold copy of the file is saved in the cloud.
- A true global file lock: Users should be able collaborate as if in the same room when working on the same files without fear of version control or accidentally overwriting a colleague’s work – even if they are located a continent away.
Only a global file system that resides in the cloud, with private or public cloud object storage serving as the new disk, addresses all of these needs and makes true collaboration possible.
In the second part of this two-article series, Andres will discuss how the cloud is enabling infrastructures to scale, connecting offices located around the world and allowing them to seamlessly collaborate around massive files.
Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Informa.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating.