Manufacturer Musings: Matt Star, Spectra Logic

Manufacturer Musings: Matt Star, Spectra Logic
Opinion
August 3rd 2016 at 1:50PM

Spectra Logic

Making the choice to migrate to the cloud can be viewed in pretty much the same way as those decisions we make in our personal life that have a significant impact: buy or lease, hire a professional or do it yourself, eat well and exercise or live life with no restraints, migrate to the cloud or stay on-premise.

When it comes to storage, the cloud can certainly seem like an attractive option for organisations incessantly generating sizable volumes of data. But moving to the cloud can be one of the most expensive decisions an IT organisation makes.

Media and entertainment companies need to consider specific key aspects of the nature and usage of their data in terms of protection, location and access before jumping on that cloud bandwagon. If these elements aren’t carefully reviewed beforehand, cloud migration could come with a higher-than-necessary price tag.

There is a compelling set of reasons to migrate data (and even applications) off premise: dynamic scaling and enterprise feature sets are just two. In the cloud model the vast majority of the burdens of infrastructure management are removed. You can get an on-demand usage model that fits most small to medium size companies. As an example, render of a movie usually only happens in the last phases of production, so a customer can stand up a render farm in the cloud only during the period they need.

But the flip side of this approach is that you no longer own the infrastructure and will be paying someone else to carry out the work of your storage administrator. Cloud also allows content sharing, but again, this is not an aspect you will be able to directly control.

For broadcast professionals, consideration of the cost incurred for access to your data and the ability to move it into and out of that system is vital. If you are thinking of storing your data in a public cloud, you also need to think about how much protection and resiliency you need. So if you want two copies it usually cost 2x a single copy. If you want that second copy geographically co-located there’s going to be additional cost attached to this, and then when you require data recovery it’s going to cost yet more again.

So the takeaway here is that using cloud when your data sets are small can work really well, but when the volume of data is sizeable, the outsourced model can prove very costly. Smaller organisations however can benefit by enjoying enterprise-class storage at lower cost than it would be on-premise.
 
The cloud approach is not new: Storage Service Providers (SSPs) have been coming and going for the last 15 years and they utilise the same disk, tape drives, and network switches as any large enterprise does, selling use of their backend hardware as a service. But what is new is the willingness for end users to adopt this model. In the past customers were worried about data ownership, SLAs and security. While those same concerns still exist today, the latest issue is now price.

On the surface cloud-based storage often looks cheap, with prices coming in at sometimes pennies per gigabyte. So how is it possible for a company to provide storage as a service so cheaply? The truth is that cloud costs much more than that; cloud providers are in business to make money, and make money they will, especially when you want your data back. One example of this is a cloud provider that charges $921 to get 10 TB out of their system

When looking to transition to a cloud-based system, customers should have a plan to also get out of that same cloud-based system. Without a back-out plan you could be mortgaging your future. So how exactly should media and entertainment professionals tackle migration to the cloud?

Firstly, building a complete cost model for deploying cloud storage will prove extremely useful. Include the cost of the pipe, in/out charges and ongoing storage charges. Weigh those against full-time equivalent (FTE), space and cooling savings. When properly modeled, cloud services usually end up saving money for the first two years; after that, the cloud cost begins to mount and over five years this can be two/five times the cost of local or private cloud storage.

Two other models worth reviewing are cloud services as the overflow and the second is to use cloud for a second or recovery copy of the data. In this case, you will greatly reduce the cost of long-term cloud deployment since in/out charges are minimised unless there is a disaster.

Private and hybrid clouds are two of the best approaches to long-term data preservation because they give lower total cost of ownership and user-driven SLAs. Private clouds can be deployed via commercially available products from many different vendors, including DDN, EMC and NetApp. There are also a number of viable open source object storage systems, including OpenStack and CEPH, that can be used in conjunction with other tools to deploy a private or hybrid cloud.

The latest innovations in deep storage gateway technology now enable the archiving of broadcast data including site-to-site replication and its management across multiple locations, using a single interface to tape and disk using cloud protocols. So, the technology to deploy a hybrid cloud archive is out there and is simpler to use than it seems.

As long as broadcast professionals don’t take a “one size fits all” approach to cloud migration, this model can be a wise choice both in technology and financial terms. With thorough planning and research into the most suitable technologies available today, cloud adoption as part of your data management strategy can be a very cost effective exercise.