Hybrid-Cloud and Multi-Cloud tend to be the direction that most organizations are heading with their cloud strategy. The problem is the path to both of these cloud strategies is full of unexpected twists and turns.
What is Hybrid-Cloud?
While these terms can and often are used interchangeably, there are some basic differences. The Hybrid-Cloud definition typically suggests that one of the locations in the strategy is the on-premises data center. In most cases, the on-premises data is the hub and the organization leverages the cloud as spokes. Hybrid-Cloud may also be Multi-Cloud in that some hybrid operations may be in one cloud and other operations are in another. The communication path between clouds though, almost always routes through the primary data center.
Use cases for Hybrid-Cloud include using it to store backup and/or archive data instead of using or expanding an on-premises secondary storage system. Organizations can also use Hybrid-Cloud for disaster recovery by replicating data to the cloud and in the event of a data center disaster, applications can start in the cloud. In the same way, organizations are using Hybrid-Cloud for cloud bursting where they move workloads, during peak times, to the cloud. Busting enables organizations to build their data center for normal day-to-day requirements, not the occasional peak.
What is Multi-Cloud?
Multi-Cloud is cross-cloud communication between, often rival, clouds. Organizations may want to move or copy data between clouds so that they can leverage another cloud’s service or take advantage of spot pricing on CPUs. The data transfer between clouds is not as big a challenge as it used to be since many providers have data centers within close proximity of each other. Multi-Cloud has given birth to a new set of solutions and companies to deliver high performance migration and unified views of multiple cloud storage systems.
The Hybrid-Cloud and Multi-Cloud Challenge
However, both Hybrid-Cloud and Multi-Cloud face a significant challenge, egress fees. Most cloud providers don’t charge a transfer fee for inbound data but they do charge fees for data leaving their cloud, no matter where that data is going. They also often charge for API calls against stored data. Charging for API calls means that even if data stays in the cloud but is subject to access from an external source, like an application running from an external service in another cloud, there is a fee.
These fees are incredibly hard to calculate and to include in the organization’s cloud operating budget. They also make moving data between clouds very expensive. Egress fees impact Hybrid-Cloud models when data is brought back on-premises and they impact Multi-Cloud operations as data is copied between cloud providers. The fees add up quickly, to the point that some organizations find it easier and more cost effective to feed multiple clouds in parallel instead of incurring egress fees. With a near real-time copy of their data in multiple clouds they can start a workload whenever and wherever they choose, but they do have to pay to store multiple copies of data across multiple clouds. While this may be less expensive than egress fees (depending on how often data is moved), it still adds unneeded cost.
The software to create effective Hybrid and Multi-Cloud strategies are now widely available and continue to improve but egress fees make using them financially less practical, thus forcing organizations to settle on one provider as often as possible. However, there are a few companies, which don’t charge for egress or API calls. Organizations that need or want to move data between clouds should look for these providers.