ktsdesign - stock.adobe.com
Cloud storage is an increasingly vital part of IT infrastructure for small businesses and enterprises alike. As such, industry analysts at IDC expect spend on cloud IT infrastructure, including storage, to reach $90.9bn by 2023.
So, what are the top cloud storage use cases? And which parts of the IT stack should you consider moving to the cloud first?
So, for now, enterprises often use hybrid technologies for secondary data applications to overcome the technical limitations – and sometimes the financial downsides – of pure-play cloud.
Storage suited to primary data in the cloud is growing, however, often in parallel with cloud-based compute. But, right now most cloud storage is object-based, which is not well-suited to primary storage. Although, as IDC analyst Andrew Smith points out, the percentage of block-based infrastructure has grown to around 30% of the market.
Top cloud storage use cases
Backing up data to the cloud immediately achieves the objective of maintaining physical separation of backup from production data and creation of a redundant, offsite copy. This “air gap” has become even more important as a result of cyber threats such as ransomware.
Cloud-based backup solutions can start as simple manual processes, such as copying key files across to Dropbox or similar consumer-focused online services. These can suffice for very small businesses or branch offices with simple requirements and datasets with few privacy or compliance concerns.
At the other end of the spectrum, enterprise systems need tiered backups, with management software pulling data from host applications via local high-speed arrays and into the cloud. This can entail the use of appliances at the network edge. An increasing number of backup tool suppliers can manage the translation from block and file to the cloud’s object-based storage formats, and back.
The downsides of cloud backups include that need for format conversion, the possible cost of data egress should you need to recover the data and the complexity of restoring data to the application’s original format. There are also compliance risks that can come with manual backup to off-the-shelf cloud storage.
There is also the question of how to create the initial backup. Organisations with large data volumes will have to consider “seeding” backups to disk, or even tape.
An emerging requirement is cloud-to-cloud backup which meets the need for businesses’ greater data security and redundancy for cloud-based applications.
Data archiving and long-term data retention in the cloud is becoming more attractive to CIOs as the costs of cloud-based storage fall. Suppliers offer near-line and cold-line storage, with services such as Amazon’s S3 Glacier and Glacier Deep Archive just two examples.
Suppliers claim cloud-based storage is significantly cheaper than tape. This point is arguable, but costs are falling rapidly. Also, cloud-based archiving offers a similar redundancy advantage to cloud-based backup. Indeed, removing the need to courier tapes off-site is a plus for the cloud.
Cloud storage is, for most practical purposes, infinite. If the cost is affordable, then there is no reason not to keep data for future use. The cloud’s automated and auditable management will also score highly with businesses with strong compliance and data retention rules.
On the downside, there is a lack of standardisation around cloud-based archiving. Although archiving solutions support a range of common application programming interfaces (APIs), including REST and OpenStack Swift, and support multiple storage suppliers, cloud-based archiving services typically cannot talk to each other.
This could make it hard to switch between suppliers. Organisations might have to restore their archives to local media first, and could be hit by data egress fees. Then there is the question of how to start the process. As with backup, unless it is a new business (or a new application), data will take a long time to seed to the cloud.
Nonetheless, expect cloud-based archiving to grow, not least because of the growth in (cloud-based) analytics and artificial intelligence (AI) tools that can put vast stores of data to new uses.
Disaster recovery in the cloud takes cloud-based backup a step further. With the right planning and resources, an organisation could conceivably replicate almost all its infrastructure to the cloud. Replicating essential tools, down to the desktop computing level, is relatively straightforward. And plenty of businesses already run core applications – from Exchange email servers, to enterprise resource planning (ERP), human resources (HR), customer relationship management (CRM) and salesforce management – on a software-as-a-service (SaaS) basis.
Moving to cloud-based storage will make disaster recovery and business continuity easier. SaaS makes it easier too. The challenge is to integrate cloud storage into the wider disaster recovery (DR) plan.
The key metrics of disaster recovery – the recovery point objective (RPO) and the recovery time objective (RTO) – do not disappear with cloud solutions.
RTO in particular can be a challenge because of bandwidth and, to a lesser extent, latency constraints. Is there sufficient bandwidth to create the data copy, and then to bring it back into a production system? CIOs might find that local backups to on-site arrays still provide the best first line of defence, with copies staged to the cloud. Cost can be a factor, with organisations not knowing exactly what recovery from the cloud will cost until they invoke it.
And, as with on-site DR, no plan will be effective unless it is put to the test.
Emerging applications: Ones to watch
As cloud storage matures technically and financially, it is becoming more suited to a wider range of applications.
File storage is one area CIOs are watching closely. Last year, Google bought Elastifile, whose file system allows customers to store NFS workloads on Google Cloud Platform. This is just one example of scale-out file storage which can span enterprise on-premise storage and storage in the cloud.
If the business is satisfied that security, privacy and compliance questions are addressed, moving file storage to the cloud is the logical next step and could provide an alternative to on-site NAS.
Some organisations use cloud bursting to move data processing to the cloud in periods of peak demand. Suppliers such as Oracle and VMware already have cloud bursting built in, so the process should be largely transparent to users and IT departments. But unless systems are optimised to work with cloud storage, there is the risk of an unacceptable performance hit.
Scientific and industrial research are areas where cloud bursting works well, as spikes in demand should be predictable. But artificial intelligence and machine learning are increasingly important to organisations in less research-driven fields.
As businesses strive to make more use of their data, cloud storage becomes a logical option. Once data is online, potentially any artificial intelligence/machine learning application could be given access to it from any location. If – and it is still an if – this decade sees AI as a practical tool for business, then AI is likely to be a service and cloud storage will serve up the data.
Read more about cloud storage
- We run through key questions to ask when specifying cloud storage, such as disk type, performance, availability and the possible hidden cost of getting data out of the cloud.
- Cloud storage offers apparent simplicity, but unexpected costs, complexity, data movement and integration challenges can arise. We look at how to avoid common cloud storage pitfalls.