By Chris Tozzi
A multi-cloud computing architecture offers several advantages: more reliability, better opportunities for optimizing costs, and an ability to address region-specific compliance concerns, to name just a few benefits.
But a multi-cloud architecture is also more expensive and difficult to maintain. For that reason, you may be hesitant to make the switch to a multi-cloud strategy without having a good reason to do so.
This article explains how to know when it makes sense to migrate from a single-cloud architecture to a multi-cloud architecture. It discusses the financial, technical and geographical considerations that make multi-cloud strategies worth the extra cost and management.
The following are the main reasons why it might make sense for an organization to adopt a multi-cloud computing architecture.
If your workloads are all relatively consistent in size and type (if you can run all of them on a basic computing instance, for example), then keeping them on a single cloud is probably the best choice from a management perspective.
Yet if you find yourself having to run many different types of workloads, a multi-cloud strategy could be a better fit, for two reasons. First, it can help you save money by running each type of workload on the cloud that is most cost-efficient. Second, you’ll be able to choose from a wider array of services, which can help you to choose the best solutions from a technical perspective.
For example, perhaps you want to use the cloud to run virtual machines, containers and Big Data analytics. These are all distinct types of workloads. One cloud provider might offer better pricing for container hosting than another, yet the situation is reversed when it comes to Big Data solutions. By having multiple clouds at your disposal, you can optimize the cost and performance of each type of workload.
Even the best-maintained cloud servers sometimes fail. It’s a fact of life.
It’s also a fact of life that one of the best strategies for maintaining uptime is spreading out your services across as wide an array of host infrastructure as possible.
Therefore, by using multiple clouds, you help to ensure that an infrastructure failure from a cyberattack that impacts one provider (such as the Dyn outage that shut down a number of major websites in 2016) will not make all of your services unavailable. They’ll still be accessible from another host if one provider is impacted.
Of course, this functionality requires you to run the same types of workloads in multiple clouds, so if availability is part of your reason for adopting a multi-cloud architecture, you should distribute your workloads accordingly.
Although all of the major public cloud providers offer hosting infrastructure in most parts of the world, some have a more extensive presence in certain regions than others. If you need to deliver content quickly to certain regions, or address legal challenges related to operating in specific regions, it can make sense to adopt multiple clouds so that you have infrastructure based in each region where you need to operate.
For example, AWS’ availability in China is limited to two regions (Beijing and Ningxia). (AWS has plans to launch a third region in Hong Kong soon.) Google Cloud has no presence in China. In contrast, Alibaba Cloud has seven deployment regions in China (on top of regions in other parts of the world). If you want to reach the China market, Alibaba Cloud can be simpler, because it doesn’t require you to work with a subsidiary (which is the case with a provider like Azure), and you don’t have to worry about slow content delivery due to host infrastructure that is not based locally.
(If you’d like to give Alibaba Cloud a try, you can take advantage of the organization’s current offer of $300 in free credits.)
There are several reasons for making the switch to a multi-cloud strategy. Don’t expect a multi-cloud architecture to be easier to maintain—In almost all cases, it will require more work. But the extra effort can pay off in the form of more cost-efficiency, higher service availability and better operations in specific geographic regions.
The Cloud and AI: A Marriage Made In Heaven For Big Data Analytics?
2,599 posts | 758 followers
FollowAlibaba Clouder - July 13, 2020
Alex - June 21, 2019
Hironobu Ohara - February 3, 2023
Alibaba Container Service - November 13, 2019
Hironobu Ohara - February 3, 2023
Alibaba Container Service - July 29, 2019
2,599 posts | 758 followers
FollowOffline SDKs for visual production, such as image segmentation, video segmentation, and character recognition, based on deep learning technologies developed by Alibaba Cloud.
Learn MoreMore Posts by Alibaba Clouder
Raja_KT February 17, 2019 at 10:11 am
You hit the nail on the head. CSPs have to address these issues somehow to garner and keep clients happy.