By Leo Leung, Vice President of OCI and Oracle Technology
The promise of cloud computing is flexible power at a reasonable price. Unfortunately for many business customers, that promise remains unfulfilled.
There are several reasons, but we’ll focus on just three of the biggest.
Computing complexity
Many cloud providers offer a bewildering array of computing instances, or virtual machines, at different price points. The vendors tout this as flexibility, but listing hundreds, even thousands, of choices is actually a great way to make price comparisons for customers horribly complicated, if not downright impossible.
To put this in context, the 451 Research Cloud Price Index (a part of S&P Global Market Intelligence) counted more than 5.9M SKUs for AWS compute. Let that sink in for a moment. Millions. It’s no wonder that many of their customers require consulting assistance to figure out which compute SKU to use.
Several providers also offer discounts for customers who lock into a given compute instance for a specified time. But, if the customer’s needs change in that period, thus requiring a new instance type, that often results in new charges.
This is why potential customers should look at the providers that offer a simple and comprehensible pricing model. They should scrutinize whether Cloud Provider A’s long-term discount is really less expensive than rivals’ options that do not require a long-term commitment.
At Oracle we argue that our bread-and-butter compute pricing is less expensive than the long term “discounts” offered by others. For a similar configuration OCI Compute costs 57% to 61% less than comparable compute options from the other hyperscaler cloud providers.
In addition, OCI offers much finer compute increments compared to other cloud providers, which typically just have pre-configured sizes. Instead of buying unneeded performance, OCI customers can fine tune their environment, further cutting costs.
An existing collaboration platform customer on OCI did a pricing exercise and found that they were saving about $225K per month by using OCI instances instead of fixed instances from another provider.
Of course, everyone needs to check pricing out for themselves, but would-be customers should be especially skeptical of long-term compute “discounts,” which may not benefit them at all.
In addition, OCI customers can add memory to an existing instance instead of changing to a new instance. Other providers often make them move to a new and bigger instance type to get more memory, which comes with unneeded CPUs. And they charge for that upgrade. That means the customer ends up paying for unneeded compute power just to get the memory they require.
This is not very cloud-like.
With OCI flexible instance types, on the other hand, customers can add memory to existing instances without incurring the expense of a compute “upgrade.”
Performance perplexity
The price of data storage, the second primary building block of cloud computing, also varies wildly between the top four hyperscale cloud vendors.
For the storage attached to cloud computing instances—the cloud equivalent to a PC’s hard drive–OCI’s nonvolatile memory express (NVMe) storage volumes cost less–between 61% and 82% less—than rivals’ similarly configured block storage options. In addition, OCI offers a simplified selection, whereas the other cloud providers field multiple generations of offerings that often have overlapping capabilities and carry significant restrictions on usage and compatibility.
Again, prospective customers should look at pricing and configurations for themselves, but the upshot is that the cloud providers that position themselves as the most cost effective are not necessarily offering the best deals.
For example, a leader in the transportation and delivery industry considered multiple cloud providers before selecting OCI. It realized that if it were as successful as planned, its resource costs on other cloud providers would skyrocket into budget-busting numbers. OCI storage provides additional capacity and performance at significantly less cost than the other providers.
Another advantage of OCI block volume storage is that customers can dial performance up or down as needed without stopping, detaching, or recreating the volumes. Other providers often tie their storage performance to the size of the volume and require a “re-sizing” process, which can mean at least a restart of the instance, if not detaching and reattaching, or even creating a brand-new volume.
Data detainment
Networking is the third major component of cloud computing, and the pricing of data transfer, or data egress, has long been a secret profit center for the other cloud providers. Cloud Provider A charges customers not only for transferring data out of its own cloud to the internet or to a customer data center but even between its own cloud regions and within its regions.
None of the hyperscalers charge for data that customers ship into their clouds. But the return trip of data from a cloud to the internet is often quite pricey. Data egress fees also tend to be unpredictable since they apply to a variety of cloud resources that show up on different parts of a cloud bill.
These charges add up and have caused many unpleasant surprises for business customers. That, in turn, makes budget planning difficult. These fees often account to 10% to 15% of a business’s total cloud spend, according to Gartner.
To be clear, virtually all providers will transfer some data out of their clouds for free before charges apply, but that limit can be laughably low. While the other providers start the meter at 100GB of data out to the internet per month, OCI customers aren’t charged till they hit 10TB per month. That’s a 100X difference. And, even after hitting that limit, Oracle’s data egress rates are noticeably lower than the other providers.
Oracle’s view is that any cloud provider can make money on networking services without assessing massive additional charges, especially as all these providers are optimizing their own networking infrastructure in order to offer better services at a lower cost to themselves.
By way of example, a leading online meeting service runs its core service on OCI to take advantage of the lower data transfer fees. This was a major attraction of OCI, especially during the pandemic when usage grew rapidly.
Summing it up
The upshot is that business customers should look at all pricing options as they continue to move workloads to the cloud. They should especially scrutinize any long-term compute “lock-in” deals before signing on the dotted line.
When it comes to storage, they should assess how easy (or not) it is to ramp storage up and down with demand.
And networking customers should take an extremely close look at data transfer (egress) prices across the board. Don’t forget data moving inside of regions and between regions, as well as outside of the cloud.
Perhaps most important, if a cloud provider presents an impenetrable price list with too many options to mention, let alone count, prospective customers should think twice, maybe three times, before proceeding.
Simple, understandable, and low pricing helps customers control their costs while moving to the cloud. And they should all focus on maximizing the return on their own investment, not on someone else’s margin.
Learn more about how your organization can save costs with OCI here.