Barry Philips is CMO and VP of Product Management at Panzura.
At risk of dating myself, I recently found a travel alarm clock in the back of my closet. My 15-year-old daughter saw it and just could not wrap her head around the concept when she has an iPhone.
While not as antiquated as a travel alarm, point-and-shoot cameras are also going the way of Dodo birds. iPhones now have great resolution and have a simple workflow to store pictures with any number of cloud providers, eliminating the need to carry a second device to take photos. I still see stopwatches when ESPN rebroadcasts NFL shows, but everyone in gyms across the country just use their smart phones. The iPhone has subsumed the functionality of these and many more devices because it is a disruptive advancement as opposed to each of these devices having separate, incremental evolutions.
We’re now seeing the cloud do the same to all forms of on-premises enterprise storage including primary storage, back-up and tape.
Analysts agree. 451 Research recently published its 2016 Enterprise Storage Outlook in which both Amazon Web Services (AWS) and Microsoft Azure become top 5 enterprise storage vendors by 2017. Further, AWS surges from sixth to second place while NetApp plummets from second to sixth. Spending on public cloud storage more than doubles between 2015 and 2017 while spending on on-premises storage falls over 17 percent.
It’s difficult for on-premises storage to compete with the cloud. Like the iPhone, the cloud is a disruptive technology that makes legacy infrastructure seem like a unwieldy point-and-shoot camera. The cloud wars have dramatically dropped the cost and features like AWS Infrequent Access reduce the cost further. Global deduplication can reduce the total storage footprint and cost even further. It’s hard to match the cloud in data protection as AWS S3 has 11 9s of durability with objects redundantly stored on multiple devices across multiple facilities. This means if you store 10,000 objects on S3, you can on average expect to incur a loss of a single object once every 10,000,000 years. To top it all off, S3 is designed to sustain the concurrent loss of data in two facilities. Very few companies can match that in their own data centers.
The scale and economics of the cloud are great if everything else is in the cloud, but how does it work for users and applications that are in offices? Users don’t want to wait for files to be downloaded from the cloud, they want the files locally with blazing fast speed all the time.
This is accomplished by putting a simple caching appliance on-premises with as much flash storage as possible. By caching appliance, I don’t mean an all-flash array. An all-flash array certainly provides the blazing speed that users need, but still requires software, hardware, and processes for DR, back-up and archiving at every site. A caching appliance relies on the cloud for durability and redundancy, but provides the local performance and enterprise features that are expected from traditional enterprise storage. Since the cloud is the “back-end” to all the sites, features like dedup can now be global in nature instead of only providing that benefit on a site-by-site basis. Of course, the caching appliance also has to be the translator between a cloud’s RESTful API and a file system interface that is used by most enterprise applications and users.
Of course, this caching device doesn't always have to be on-premises. The caching device can also run just as easily inside the cloud. This makes a few things possible. First, applications that were developed for the corporate data center can now run without a single change in the cloud. Additionally, if an office is close enough to a cloud data center with respect to latency, that office could access a cloud-based caching appliance over a network connection. This eliminates any on-premises storage infrastructure, but still provides enterprise-grade storage performance and features to that office.
Data storage is expected to grow 800 percent over the next five years, and 80 percent of new data will be unstructured data. The trouble is, 70 percent of unstructured data has not been accessed in 60 days, and 90 percent has not been accessed in six months. This requires a more modern approach to storage architecture, and the cloud is the key to this new architecture. '
The cloud doesn’t require the overprovisioning of storage capacity to achieve performance, overspending on expensive storage media for inactive data, or the overbuilding of data centers to house increasing amounts of storage infrastructure – yet still provides the performance and features that users require.
There is no need to worry about data growth and capacity/forecasting ahead of that growth. Spending transforms from a model where you pay upfront for capacity to grow into to a model where you pay as you grow. There’s no separate software, hardware, or processes for DR, backup, or archiving as these are just part of the solution. Most importantly, IT can significantly reduce the time to manage storage as storage now becomes a centrally managed service.
Technology inflection points come and go, but there are few that are as transformative as the cloud. The cloud has already disrupted the way applications are deployed and developed with a combination of SaaS and PaaS. The next step of cloud disruption will completely change the way IT infrastructure, like enterprise storage, is procured, consumed, and managed. In just a few short years, new hire college graduates in IT will have the same reaction to tape as my daughter did to my archaic travel alarm.