When Does a Multi Cloud Strategy Make Sense?
On February 28th, 2017, the Amazon Web Services Simple Storage Service (AWS S3) in the North Eastern United States went dark, taking large chunks of the cloud-based Internet services with it. Multi-cloud deployments leverage virtualization to remedy the risk of downtime and act as a force multiplier that provides additional benefits.
The result of the S3 outage was that some of the largest SaaS providers on the web went offline, victims included The AV Club, Trello, Quora, IFTTT, and websites created with Wix. While this may have been a black swan event, never to repeat, the disruption for subscribers that depend on Amazon for compute capabilities could have cost them more than lost revenue.
When A Multi-Cloud Downtime Mitigation Strategy Makes Sense
Multi-cloud deployments use parallel subscriptions with two or more cloud-based infrastructure service providers. In essence, a multi-cloud deployment is a variation of a hybrid cloud that combines subscriptions of cloud service providers. The principle is that duplication diminishes downtime risk by shifting traffic to unaffected infrastructure and platforms service providers.
The leading cloud services providers like AWS, Oracle Cloud, IBM, Google Compute and Microsoft Azure all promise minimal downtime in their service level agreements for cloud-based services. It takes exceptional requirements for small and medium-sized companies and startups to justify subscribing to more than a single vendor; a few minutes of downtime a month may not be fatal.
When Software Roams Free
Multi-cloud strategies and the software that enables it are all part of the continuing trend of software-defined data centers. The higher levels of platform and software in the Cloud can ultimately be divorced completely from specific hardware infrastructure.
Any multi cloud solution must balance the benefits with the additional overhead costs that it incurs. It is not just the cost of duplicate subscriptions but monitoring and VPN connections as well, for example, where the performance improvements will override the shortcomings and costs.
Matching applications to the cloud that best serves them to boost performance past the breakeven point that justifies the decision to adopt a multi cloud deployment. Multi cloud allows companies to match workloads to the services best suited to the tasks. Examples would be IBM infrastructure running on IBM Cloud or Microsoft Dynamics ERP running on Azure Cloud.
A Non-Linear Cost for Small Business
There is always a possibility that an anomaly could extend to an entire class of infrastructure. A multi-cloud strategy combined with a commitment to internal hyper-converged infrastructure would increase redundancy significantly, but at the cost of raising operation overhead costs too.
There will also be consideration of issues such as security; the more accounts that share data, the greater the attack surface for cybercriminals. Multi-cloud solutions are not seamless unless they can align features to capabilities; the compute resources of different providers are likely to handle data in subtly distinctive ways, which then accumulate errors exponentially.
Balancing Exceptional Considerations
However, if applications do require the collaborations of disparate vendors, the answer is often available in a software-defined solution. A hypothetical situation might arise where testing and development runs best on VMWare vSphere, but the production environment works on the Google Compute KVM hypervisor. To achieve a successful deployment might necessitate additional tools like Oracle Ravello to overcome interoperability issues.
For now, multi-cloud strategies remain enterprise solutions where the direct cost of lost minutes of downtime outweighs the cost of multiple deployments. However, SMB SaaS providers should consider these strategies when the cost of unlikely events become significant factors that could destroy their businesses.
Multi-cloud deployments make sense as part of disaster recovery plans. The potential to create excellent and efficient software innovations are much more compelling reasons though. By selecting the finest resources that the Cloud has to offer, small and medium-sized digital businesses can create the agile software applications that will make them the enterprise scale services of the future.