How to escape hyperscaler transfer fees
chri6103
Thu, 09/30/2021 - 05:05
How to escape hyperscaler transfer fees
September 30, 2021
by Michael Levy, Director of Product, Storage and Colocation, Rackspace Technology
Hyperscale cloud providers dont charge you for importing data to their platforms. But they do hit you with significant fees when data is removed. These egress fees can be cost prohibitive for many companies. In fact, according to IDC, 99% of U.S. technology executives said they incur planned or unplanned egress fees at least on an annual basis.
If I were in charge of a companys IT budget, I would want to use the hyperscale cloud. In fact, Id likely want to leverage every major hyperscaler, because each one offers unique proprietary toolsets that could potentially help drive significant business value for my company. However, I would be concerned that Id be taken for a ride every time a workload moves from one cloud to another, and results in huge egress fees an outcome that would undoubtedly land me in big trouble with my CFO.
One lower egress charge option
There are several editorials lambasting hyperscalers for charging for data egress, but not for ingress. Some have even condemned it as a predatory practice. However, I do give hyperscalers the benefit of the doubt. It takes a lot of internet connectivity to transport information back to customers around the world and that comes with a monumental price tag. To be fair, they do offer a pathway for lowering egress charges significantly by directly connecting their platform to your private network.
For instance, AWSs egress charges start at $0.09/GB. If you ride a private network into one of their Availability Zones (AZ) directly, that initial rate goes down to $0.02/GB. For Google Cloud Platform, the discount is even greater, with rates dropping from $0.12/GB to $0.01/GB. These are generous discounts, and they would generate respectable savings for customers with a lot of outbound traffic. The challenge, however, is around how to establish these connections.
Challenges of making a direct connection
To link up directly with a hyperscaler on your own, you have two options. The first is calling your local telephone company and asking for a private MPLS or ethernet connection from your premises to the hyperscalers on-ramps. This is expensive and can take months to implement. Unless youre within a few miles of the data center that hosts the on-ramps, youre going to introduce latency into your network, which would tarnish your user experience.
Your second option is to establish a physical presence at the colocation data centers where your hyperscalers are located and interconnect your gear with their network. Within a given metro area, there tends to be one or two colocation data centers where all the hyperscalers amass. The high-profile residents drive up these colocation sites value its as if theyre the beach-front property of the internet infrastructure world. Not only are monthly costs to deploy your hardware jacked-up tremendously, but the fee for a fiber cable between you and each hyperscaler increases too often to the tune of hundreds of dollars per month, per wire. Thats on top of the tens of thousands of dollars in capital investment you must endure to procure the network and storage hardware required to receive data leaving the hyperscaler. Finally, youll need to find and hire experienced engineers, a commodity that is in short supply.
Introducing Rackspace Technology Data Freedom
Rackspace Technology can save you from all of that. We have the network equivalent of coastal villas at almost all the coveted pieces of digital real estate. Theyre all tied together with our private backbone and directly connected with both the major hyperscalers and the leading cloud ecosystems. This allows us to transport customers to almost any service provider and edge location without touching the public internet even once.
And because Rackspace Data Freedom users can avoid large-scale data migrations to the public cloud, they can achieve transit cost savings of up to 83% by using the Rackspace Technology private network instead of the public internet.
Rackspace Technology has decided that its time to share our savings with our customers via a new platform called Rackspace Data Freedom. The platform rolls up storage and low-latency direct connectivity to all hyperscalers into a single solution for a fixed, predictable monthly cost. This spares you from out-of-control egress fees by allowing you to repatriate your hyperscale storage into a private environment.
And because Rackspace Data Freedom users can avoid large-scale data migrations to the public cloud, they can achieve transit cost savings of up to 83% by using the Rackspace Technology private network instead of the public internet.
Declare your independence from out-of-control egress fees
When data exits a hyperscaler to the disaggregated storage managed by Rackspace Technology, you pay the reduced rate because youre using our private network. Whats more, we weave all of your endpoints together, so all traffic bouncing between them circumvents the public internet and, instead, uses our private backbone. So, traffic going between, say, Google Cloud and AWS benefits from lower egress charges, whether it touches the Rackspace Technology managed storage environment or not.
To get the most out of Data Freedom, our customers run their applications to callback only the information that must be pulled into hyperscale compute and native toolsets. Often, less than 5% of data within a workload is hot, meaning it must make its way over to the application for processing to complete the task at hand. As a result, the bulk of the data wont even face data transit fees at all, while the remainder is charged at significantly discounted egress rates.
The practice of disaggregating storage from public cloud compute is often dubbed data independence. With Data Freedom, youre not just liberated from exorbitant bills, youre also unburdened by the threat of future egress charges when considering multicloud strategies. With this major inhibitor removed, you gain the liberty to have each of your workloads deployed on the infrastructure platform or cloud-native toolset that best meets its needs. This is the missing link for companies whove wanted multicloud freedom, but couldnt afford it.