Enterprise cloud storage is now very much the default – for backup and archives, that is.
If you are using the cloud to store backup and archives, then welcome to the club. This cloud usage case is widely popular with reason: 80% or so of corporate data is inactive data. The cloud is an obvious choice to store this growing volume of data, whether to nearline services with moderate expectation of recovery or to cold storage with minimal recovery needs.
It’s not as common for the enterprise to use the cloud with production data. Clearly cloud storage is not yet fast enough for high performance applications. High I/O and low latency still need LAN speeds and on-premise high performance arrays.
But there are growing usage cases for storing primary data in the cloud. These applications do not need high performance as much as they need high capacity, on-demand environments. Instead of trying to provision the on-premise data center to suit big processing spikes, the cloud acts as an extension to the production environment. This produces scalability far above what most data centers are capable of. Extending this type of primary storage to the cloud saves on capital purchasing costs, licensing and maintenance costs, data center real estate, and IT management time. At least as importantly, it accelerates business value by providing massive resources to critical business applications.
In this article we’ll discuss extending production environments to the hyperscaled cloud and how to get there.
Issue with Enterprise Cloud Storage
First let’s define primary storage. In its simplest terms, primary data actively interacts with its application. Creation, modification, reading, writing: each action requires an acceptable level of performance and data is at a higher value than most inactive data. Enterprise concerns over cloud computing performance and security have retarded efforts to move primary data to the cloud.
But the times are changing both on-premise and in the cloud. Cloud providers are dedicated to making their offerings faster, more secure, and higher value with inline dedupe and compression, cloud-based snapshots, geo-location, and accelerating data traffic. And IT’s growing comfort with the cloud as a storage target is paving the way to pilot projects for storing primary data in the cloud.
Now let’s make the case for deploying primary data in the enterprise.
Making the Case for the Enterprise Cloud
Cost and time savings is the main reason that IT stores inactive data to the cloud. The business case is simple to make as long as security and compliance are in place.
However, the argument morphs when it comes to extending primary data to the cloud, using it as an active production tier in concert with on-premise infrastructure. This level of service is more expensive than simply storing inactive data to cloud-based cold storage. And if all you are doing is simply moving a few application servers to the cloud, the investment may not be worth the return.
The business case for primary storage in the cloud is not about saving money on the status quo, but adopting high-value business applications that would otherwise overwhelm on-premise resources.
Specific cases for hyperscaled cloud environments include Hadoop-as-a-Service, where your cloud provider builds your Hadoop clusters online. Another use case is a large construction company that uses the cloud to version and deliver digital blueprints to job sites and mobile devices all over the world. Another active use case is hugely scaled rendering farms for animation and visual effects. One major visual effects company has 35 high-performance rendering servers on-premise – and has used up to 35,000 rendering cores in the cloud using Google Cloud Platform Zync Render.
The simplest way to accomplish this is to invest in a system that does this for you: a cloud-enabled storage system that treats on-premise and cloud deployments as a single extensible infrastructure.
Be certain to invest in a system that is scalable on the on-premise site as well as the cloud site. You do not want get stuck with a single appliance that cannot communicate with its counterparts in remote offices. Besides Google Storage Cloud, systems built for active cloud processing include Microsoft Azure StorSimple, CTERA Cloud Storage Gateway and AWS Storage Gateway. Panzura and Nasuni also offer NAS/single file system solutions using hyperscaled clouds.
Cloud Storage, Corporate Will and the CIO
To paraphrase The X-Files, the technology is out there. But what about its adoption? It seems to be an easy on-ramp to the cloud, and most yearly CIO surveys report that the CIOs plan to spend more on cloud services. And they may, but those services are usually simple value-adds like more cold storage.
Primary storage in the cloud is a bigger scope. That usage case concentrates on high value IT projects made possible by the cloud instead of merely presenting it as a low-cost alternative to local storage. Selling this business use case takes committed IT and a CIO who can communicate the strategy’s value to senior managers.
IT must be committed because they’re the ones who will have to take the time and resources out of already busy jobs to research and test cloud appliances. There are also the people who will need to convince end users that their data is safe, and demonstrate that the new way of doing things beats the old one by a long shot. IT will also need to communicate success to the CIO, who in turn communicates that success to senior management.
The good news is that a test project may be so successful that end-users – including senior executives – may begin to request new services even before they’re ready to roll out. This type of gateway project becomes an attractive on-ramp to the cloud because it demonstrates its usefulness to larger numbers of people.
Enterprises are rarely early adopters, but primary storage in the cloud is not an incipient technology. It is a maturing one with room for growth and development. The technology is in place now, and often makes more sense than building out on-premise storage for massive data processing or large-scale file sharing. Understanding how the cloud serves these business applications is the key to success for IT and the business units, and visible success will drive even larger cloud adoption projects.