The rise of the cloud has had a massive impact on data protection, making backup processes almost unrecognisable from just a decade ago. The cloud is increasingly popular as a site for production workloads and their backups, while physical and virtual servers on-site decline.
Meanwhile, disaster recovery (DR) using the cloud is in widespread use, despite some challenges. And native cloud-based backup of software-as-a-service (SaaS) platforms such as Microsoft Office 365 is largely untrusted.
Those are some of the findings of the 2021 Veeam cloud protection trends report, which questioned 1,551 IT decision-makers in 14 countries about data protection and the cloud.
The most general finding of the survey is that the cloud as a location for data protection is increasing hugely, especially since before the pandemic.
According to respondents’ estimates, use of physical servers in the customer datacentre will decline from 38% of the organisation’s data in 2020 pre-Covid to 24% in 2023.
Meanwhile, use of virtual machines in the datacentre will decline from 30% in 2020 to 24% in 2023.
But use of virtual machines in the cloud is set to increase from 32% in 2020 to 52% in 2023.
In keeping with that finding, the cloud is now a mainstream location for high priority and normal production workloads for a majority of respondents (47% and 55% respectively). One-fifth (21%) use the cloud as a secondary site for DR and 36% use it for development.
Despite talk of cloud repatriation – bringing workloads back from the cloud to the customer datacentre – this mostly happens to those that have been developed in the cloud but for use on-prem (58% of those questioned had done this).
Only 7% had had second thoughts and repatriated cloud workloads back in-house. About one-quarter (23%) had brought workloads back on-site after failing over to the cloud during a disaster.
Data protection strategy in the cloud is increasingly not handled by the data protection team in the IT department. Only about 33% of those questioned said this was how they do things, with “central IT”, the cloud decision-making team and application owners more likely to be involved.
Cloud and disaster recovery: A good fit
Use of the cloud as a DR and secondary data location is well established, with 40% reporting its use for these purposes. Only one-fifth (19%) said they do not use any cloud services as part of their DR strategy.
For more of those (40%), data is mountable in the cloud but run from the customer location. For 25% of respondents, data has to be pulled back from the cloud first. About one-eighth (12%) are fully cloud-based in their ability to spin up servers and start work again.
Despite DR being a good choice as a cloud deployment, there are challenges. Hosting restored servers that were in one location and bringing them back up elsewhere can be fraught with problems, including how to reconnect networks while ensuring they are secure. If there is a mix of cloud and on-prem, the difficulties can be multiplied.
Key challenges in cloud DR identified by those questioned included network configuration (54%), connecting users in the office (47%), securing the remote site (43%) and connecting home workers (42%).
For those not using the cloud for DR, key concerns are security (20%), already using a third-party DR location (18%), cloud infrastructure being too expensive (14%), existing use of multiple datacentres for data protection (14%) and lack of manageability in cloud DR (12%).
Native Office 365 data protection: Not trusted
The Veeam survey also asked specifically about Office 365 and found that about one-third (37%) of respondents use backup other than that provided by native features, so-called cloud-to-cloud backup.
Key reasons given were to protect against accidental deletion of data (54%), against cyber attack (52%), internal threats (45%), to provide better restore functionality than in-built capabilities (45%) and to meet compliance requirements (36%).
Containers
Finally, when it came to protecting data used in containerised applications, the largest number of respondents (37%) said stateful data was protected separately and backed up in that location, possibly indicating that it is held in dedicated local or shared storage, such as an array.
Meanwhile, 19% said their containerised applications’ data did not need to be backed up, and 28% said their container architecture is natively durable.
Only 7% use a third-party backup tool to protected containers’ stateful data, while 7% do not back up container data and are looking for a solution.
