Virtualization Technology News and Information
SIOS Technology 2022 Predictions: DR will fuel the move to multi-cloud

vmblog predictions 2022 

Industry executives and experts share their predictions for 2022.  Read them in this 14th annual series exclusive.

DR will fuel the move to multi-cloud

By Cassius Rhue, VP, Customer Experience at SIOS Technology

While cloud computing will remain a core component of many an organization's IT strategy, 2022 will see many companies expanding beyond a single-cloud strategy. Not surprisingly, companies making a first foray into the cloud have tended to standardize on a single cloud vendor. Amazon Web Services Elastic Cloud (AWS EC2) has maintained a first-to-market leadership position in the world of infrastructure as a service (IaaS), taking in 31% of the worldwide cloud infrastructure spend in Q2 of 2021. Organizations that were predisposed to Microsoft have generally gravitated to Azure, which took in 22% of the worldwide cloud infrastructure spend in 2021. Google Cloud Platform (GCP) took in 8% of the worldwide cloud infrastructure revenue, but in 2022 we predict a greater adoption of GCP in two particular use cases: With more cloud experience, companies running on AWS and Azure will expand beyond single-cloud standardization and increasingly add Google to their cloud computing mix in order to take advantage of Google's leading AI, ML, and complex analytics tools. Furthermore, companies looking for cloud-based disaster recovery (DR) options will consider enhancing their DR protection for their cloud-based high availability (HA) configurations by replicating to a different cloud.

Overcoming the challenges of multi-cloud for application specialization

Many companies have found themselves accidental adopters of multi-cloud strategies. In some cases, a multi-cloud strategy is the by-product of growth through acquisition. In others, independent decisions were made in decentralized IT departments and suddenly applications were running in different clouds to take advantage of cloud vendor-specific features. GCP, for example, offers features that favor analytics workloads so it made senses to run those workloads there. But in both cases the experience has also been enlightening: Organizations are learning that a multi-cloud approach creates new options: They can choose different cloud offerings to meet individual application needs, operational demands, and high availability requirements.

Regardless of whether multi-cloud is deliberate or accidental, IT teams want to protect applications from downtime and data loss without adding complexity. While AWS, Azure, and GCP all provided high availability (HA) for their infrastructure - meaning that they will guarantee a service level agreement (SLA) of 99.94% uptime for their virtual machines (VM) - they do not guarantee application-level HA. That means that a VM could technically be up and SLA-compliant but the organization may be unable to run their critical applications. In several cases, human error has disrupted service to applications in cloud environments, even when cluster nodes were located in different availability zones (AZs).

To overcome this vulnerability, organizations have been adopting a strategy of choosing an HA vendor that can not only ensure proper application and data replication within a cloud but also one that can provide HA and DR solutions capable of supporting all the use cases required. That is, they can automatically replicate applications and data from Windows and Linux OS systems running on AWS, Azure, and GCP.

For organizations that simply cannot afford to have their critical applications and data go offline in the event of a whole-cloud failure or regional disaster, a multi-cloud DR solution makes sense.

Overcoming the challenges of multi-cloud for DR

Given the expense and the complexity of building out a multi-cloud DR solution, organizations will be selective about the applications and data they support with a multi-cloud DR solution. They will try to provision a DR infrastructure that mirrors - as closely as possible - the primary cloud infrastructure supporting their critical applications and data. The challenge for most companies will actually lie in the selection and configuration of the DR services that will keep the two cloud infrastructures synchronized and that will facilitate failover between the cloud environments if a disaster takes down or threatens the primary cloud environment.

Rather than relying on backups or log-shipping approaches for a DR solution - which would require IT to source and configure replacement servers before restoring lost applications and databases - organizations will configure a DR solution that can replicate critical applications and data across cloud infrastructures. In the event of a whole cloud catastrophe, they will have a replicated copy of critical application(s) and data in a different cloud DR.

Some companies will choose to configure a multi-cloud DR solution in the same region that houses their primary cloud infrastructure, potentially enabling the organization to use a DR solution's high speed synchronous file transfer services to replicate data from the one cloud to the other. More likely, given the possibility of slower inter-cloud communications, the DR solution will be configured to use asynchronous data transfers to replicate production data to the secondary cloud infrastructure. Asynchronous data replication will be slower than synchronous data replication and possibly create a situation in which the data in the DR cloud is a few seconds out of sync with the data in the production cloud, but that risk will be weighed against the added benefit of building out DR infrastructure in a geographically distinct region: This approach will ensure that whatever brings down the primary cloud infrastructure - particularly if the disaster has its roots in a regional event - will not also bring down the secondary cloud infrastructure they'd been banking on using in an emergency.

 # #


Cassius Rhue 

Cassius Rhue leads the Customer Experience team at SIOS Technology responsible for customer success spanning pre-sales, post-sales and professional services engagements. With more than 20 years of experience at SIOS and a keen focus on the customer, Cassius' significant skills, and deep knowledge in software engineering, development, design, and deployment specifically in the HA/DR space are instrumental in addressing customer issues and driving their success. Cassius has a Bachelor of Science (BS) degree in Computer Engineering from the University of South Carolina - Columbia, SC.

Published Friday, December 03, 2021 7:34 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<December 2021>