My rant for the day... REST vs RESTful

Disclaimer...  Getting on my SOAPbox...

Those that know me well know that - in certain areas like technology - I pay close attention to detail.  Some of my closest friends have even gone as far as calling me borderline obsessive when it comes to paying close attention to detail.  This certainly doesn't extend to other areas (you'll probably find spelling or grammatical errors in this blog post, for instance), but one that has really annoyed me lately is the loose use of the term 'REST' by many in the industry.

REST, or Representational State Transfer is an architecture and characterizes behavior, specifically that clients and servers exchange requests based on the representation of an object, typically its current state.  In terms of cloud storage, REST generally means PUTting or GETting the latest copy of an object based on some supplied metadata and in its simplest form is nothing more than a query on top of HTTP.

But some, when discussing how cloud storage is accessed, say "cloud storage is accessed using REST".

Hence my post.  REST is NOT a protocol.  It is an architecture and a behavior.  HTTP is a protocol, and the way it is typically used with cloud storage conforms to REST as an architecture and a behavior, thus the interface used for accessing cloud storage is RESTful.  But, the protocol is still HTTP.

The correct way to say this would be "cloud storage is accessed using a RESTful HTTP-based interface (or API for that matter)".

Off my SOAPbox (ha) and on to breakfast.  Have a wonderful day everyone.

How the cloud is transforming data storage architectures

It is absolutely amazing to be on the forefront of a paradigm shift in the data center.  The last five years have spawned the emergence of the public cloud, and we've all had the opportunity to witness its ongoing process of maturation.  My feeling is that we haven't even seen the full tip of the iceberg at this point, and that the possibilities for how cloud architectures can revolutionize the way we live and how businesses are managed are endless.

We're seeing such transformation now in customer data center storage architectures.  Most of the customers we speak with have what I would generously call a "legacy" architecture comprised of a series of (mostly) disjoint products, technologies, management stations, and certainly workflows.  Simple spikes of, and the ongoing continual growth of, data with such an architecture leads to enormous complexity and cost, and forces most buyers to isolate business-critical considerations such as disaster recovery and long-term content archival to only those applications that need it most and those where the costs can be justified by the net impact to the business should the information not be available.  Put simply, most customers are realizing that:

• Their infrastructure is currently or has already grown beyond their control
• They are unable to cope with their data growth, leading to lower levels of service quality and SLAs for less-than-critical applications (and the point of demarcation keeps moving upward toward more expensive applications that are core to the business)
• They can't keep all of the data they want to, because it is too costly and there are too few human resources to do so and inadequate technologies to do so
• There must be a better way...

While I've never been an advocate of saying that the cloud "solves every business problem", the number of problems it *does* solve very well continue to increase.  The emergence of hybrid cloud storage solutions (such as the solution provided by my company, StorSimple) have allowed customers to manage storage infrastructure in a familiar manner while collapsing the number of infrastructure components needed for the end-to-end lifecycle of their data - and remove many costly capital and operational cost elements.  Such devices are deployed in the data center, managed just like existing storage arrays, but the key difference is that one of the "disk trays" behind the device is a connection to one or more public or private storage clouds.  This architecture has - so far - proved itself very effective for applications that are centered around user data, virtualization, and long-term content retention.

Within these devices, intelligent algorithms that manage and monitor data access patterns - coupled with deduplication - allow these systems to dynamically and transparently move data across tiers of storage within the data center appliance and the cloud to ensure that the right blend of cost and performance are achieved.  Given that the cloud "disk tray" can be tens, hundreds, or thousands of miles away, off-site backups become integrated with your primary storage management tools, and data protection workflows are reduced to simple snapshot management - something never before possible with traditional arrays that have to allocate expensive storage capacity for point-in-time backups that can now be stored reliably and inexpensively in the cloud.  Another unique benefit our customers have mentioned is that all of their backups stored in the cloud are portable, meaning they are reachable from any location in the world with an Internet connection and a hybrid storage device.  This means that the line of demarcation that customers have to draw when determining which applications are appropriate for building costly disaster recovery architectures can shift downward to cover a larger portion of the business infrastructure because of the reduced cost and accessibility of their backups.  From an operational standpoint, these devices allow customers to deploy a single, unified solution that provides the functionality of:

• Primary storage devices
• Disk-based backup storage devices with deduplication
• Archival storage systems
• Tape-based backup storage systems
• Secondary facilities for disaster recovery (using the cloud)

As I mentioned earlier, this may just be the beginning.  Hybrid storage devices are providing a level of convergence of storage management capabilities that may possibility be the most significant that the data storage industry has ever seen.  Taking advantage of the cloud for both primary and secondary data storage when leveraging the intelligence of a hybrid storage solution radically reduces complexity and cost while allowing customers to better manage their data lifecycle, improve data protection while broadening the scope of disaster recovery coverage, and implement forward-thinking value-centric I/T services that move their business forward.

I'd be interested in your take on what the evolution of the cloud will provide businesses beyond what we're already seeing today.  

Cheers,
Joel
@jchristn
@StorSimple