The shift of enterprise IT to cloud-based services continues to grow unabated and has accelerated during the pandemic. Consumption of public cloud services, such as Google Cloud, are growing collectively at a 18% CAGR according to a recent report from Omdia.

Enterprises are migrating to the cloud, not just due to lower costs, but due to greatly increased flexibility. The ability to scale up IT infrastructure during periods of peak demand and/or growth and pay for it on an on-demand basis is very attractive. An additional benefit is the ability to divide their cloud spend across one or more cloud providers and/or shift their spend from one to another is referred to as a multi-cloud model as shown in Figure 1.

Figure 1: Enterprise to multi-cloud edge

To facilitate the migration of enterprise IT workloads to the cloud, there is a requirement for higher speed connections from the enterprise edge to cloud provider that are scalable with enhanced security to best protect critical business data. Shared IP network connections to the cloud are acceptable for lower speed (10Gb/s) connections and below. However, when secure, higher speed connections are required to the cloud, connectivity via the IP network can become overly complex, expensive, and inefficient when compared to the optical network (Optical Fast Lane) that can provide a more efficient, cost-effective, and secure option for enterprises needing to reduce their workload migration times to support their evolving business objectives.

For the multi-cloud market to succeed, it must reduce the friction for enterprises to migrate their workloads to a cloud provider, as well as between cloud providers – on demand. This is analogous to the days when you had a mobile plan with one carrier, and to switch to another carrier, you had to switch mobile numbers, which was too complex for most customers, so they stuck with their existing carrier. Only when consumers could keep their phone number when they switched carriers (through Local Number Portability), did it make the mobile market truly competitive leading to improved choice, pricing, and innovation. This is what we’re trying to achieve in the multi-cloud market.

Figure 2: Cloud transfer bottlenecks for large datasets

A key area of friction for the adoption of multi-cloud services is faster workload to cloud migration times. Whether migrating data sets from an enterprise to a cloud provider or from one cloud provider to another, transferring 1 Petabyte (PB) of data at 10 Gb/s, will take 12 days as shown in Figure 2, which is unacceptable. Some cloud providers have even resorted to physical transport (e.g., 18-wheeler trucks equipped with hard drives) to address this connectivity gap.

The vision is to scale cloud connectivity so that workload migrations to the cloud would require hours instead of days. In the above example, transferring that same 1 PB at 400 Gb/s would reduce data migration time to 6 hours. 

The vision of the multi-cloud market is where connectivity to the cloud is aligned with the elastic consumption model of the cloud. This is where an enterprise can request more capacity from the network to migrate a very large data set rapidly, then network capacity can be freed up for other enterprises to use. 

What are the Requirements for Optical Fast Lane to Multi-Cloud?

• There is a need for direct enterprise-to-cloud connectivity that can scale from 100Gb/s to 800Gb/s on demand to reduce cloud migration times from days to hours.
• Connectivity to the cloud must be simple to request and manage, extremely reliable and secure–using tools that do not impact network performance.