Driving value from cloud infrastructure and technologies is now a key part of most business strategies, evidently high on the agenda across nearly all industries. In particular, a hybrid cloud approach, combining public and private clouds with on-premise and edge, is increasingly valuable. Staff and consumers alike rely heavily on the cloud to complete day-to-day duties and activities, which can lead to resources becoming strained by demand for service. An increasingly popular way to maintain performance and mitigate this challenge is by shifting workloads into a separate cloud environment – a process known as cloud bursting.
Here, we explore what cloud bursting entails, and how businesses can utilise this application deployment technique effectively, while mitigating risks.
Best cloud management platforms for your business — With cloud management being paramount to most businesses, here are the platforms that are driving the most value on the market.
How it works
Cloud bursts occur when applications running in private or on-premise environments (e.g. a data centre) enter a public cloud to utilise further resources. In the case in which resource demand for on-premise or private cloud-based services reaches a peak, businesses can configure two-way access to public cloud resources in order to keep operations running. This can be put in place manually or automatically, depending on the company’s preference and needs.
“For example, private cloud data processing activities can reach peak demand quickly. Having the capability to scale out into cloud services during those peak times allows the operation to continue with the benefits of being able to scale back when needed,” said Jamie Shields, lead cloud architect at web and mobile product development firm xDesign.
Cloud bursting is valuable for organisations that oversee high variations of data workloads, or a predictably high temporary load.
According to Rahul Pradhan, vice-president of product and strategy – modern transactional data stack at NoSQL cloud database provider Couchbase, two main areas of applications can especially benefit. One is short-lived, highly intensive seasonal peak applications, seen in sectors such as retail or e-commerce. “Think Black Friday or peak holiday booking seasons, where loads can reach 100 times a regular day.”
The second, says Pradhan, entails any occasionally used analytics or machine learning workloads. “While not business critical, the compute capacity needed to run these far exceeds what is available in the private clouds or data centres and would be too expensive to have permanently available.
“This is where cloud bursting can help – managing the short occasions when machine learning or analytics is vital to business needs, before being shut down again.”
Manual, automated and hybrid bursting
Different types of cloud bursting approaches that can be carried out include:
- Manual bursting: The load balancer, which distributes traffic and workloads, notifies the cloud operator, leading to creation of deployments on a temporary basis for a specific purpose.
- Automated bursting: Infrastructure-as-a-service (IaaS) allows for configuration of cloud resources on demand, in line with evolving business needs.
- Distributed load balancing: Workloads are balanced and monitored between the public cloud and the data centre.
Considering these types of cloud bursting, Greg Adams, regional vice-president UK&I at observability and security platform Dynatrace, comments: “Manual bursting enables organisations to provision and de-provision public cloud services at their own discretion, based on alerts that have been created to warn teams of an emerging problem.
“This works well for organisations that want to leverage public cloud resources for specific moments that they can anticipate in advance, but comes with an increased risk of human error that lead to inefficient spend or missed problems that impact user experience.
“Other businesses will prefer to go down the automated route, which utilises tools to monitor resource consumption and automatically spin-up and down additional compute power in the public cloud as demand rises and falls.”
Elsewhere, firms may opt for a hybrid approach, permanently distributing workloads between on-premises and public cloud infrastructure. Here, the public cloud is designed to scale up from minimum capacity when demand increases.
Speed of the essence for data centre compliance — The larger your business, the more important it is to consider data centre compliance. A non-compliant data centre could lead to a breach of data protection and privacy regulations.
Key benefits
When exactly to put this measure in place should be thoroughly discussed between data and software development leaders, and business executives, before implementing it the cloud strategy. If implemented properly, an array of perks helping towards returns on investment can come to fruition.
“Cloud bursting is positioned as a cloud computing model that will give you benefits in cost efficiency, scalability, improved performance, flexibility and business continuity,” said Rick Boyce, managing director of cloud engineering at digital consultancy AND Digital.
“There is no denying that this could be true for certain organisations. It is vital that businesses review cloud bursting options on a case-by-case basis as the level of maturity of their cloud environment is the main consideration for exploring these options.”
Through using public cloud resources to keep services running, businesses only need to pay for the cloud resources they use, while avoiding fixed capital expenditure (capex) overheads.
Hyperscaler environment differences
Key cloud bursting differences can be seen between the three main hyperscalers on the market: AWS, Google Cloud and Microsoft Azure. As with which kind of cloud bursting to implement, organisations need to keep these in mind to ensure that the approach aligns properly with specific requirements.
According to Dominik Birgelen, CEO of everything-as-a-service (XaaS) cloud platform oneclick: “In terms of offerings, AWS has a wide range of services and is often considered more mature; Azure integrates well with Microsoft products; and GCP is known for its innovative offerings.
“Meanwhile, Azure has strong integration with Windows environments, AWS offers a variety of tools for different use cases, and GCP is popular among organisations looking for advanced machine learning and data analytics capabilities.
“Pricing structures and models vary among providers, making it important to analyse cost implications for different cloud-bursting scenarios.”
Three myths about cloud costs, debunked — Here are some of the biggest misconceptions around cloud cost optimisation.
Drawbacks to address
For all the business advantages that cloud bursting can bring, the process is anything but risk-free, and many of the benefits mentioned earlier could disappear as a result of insufficient planning or forecasting of regular processes.
Many of the hazards that can arise pertain to consistency, latency and gravity problems, which can lead to increased costs. “Synchronising data between on-premises and public cloud environments can lead to issues with data consistency, latency, and potential data loss. Ensuring data integrity and accuracy across all environments can be challenging and costly,” said Arash Ghazanfari, UK CTO at Dell Technologies.
“Cloud bursting could cause critical, non-disposable data to be created in the public cloud environment into which the organisation is bursting, causing the overall data gravity to shift in favour of the public cloud environment. This can dramatically increase the cost of exit. It may make it very difficult to scale the public cloud deployment back, which can have unintended adverse budgetary consequences for the organisation.”
To mitigate cost management challenges coming due to increased resource usage, Ghazanfari advises: “Carefully monitor and manage resource usage to avoid unnecessary expenses. Effective governance and control is essential; however, managing resources and workloads across multiple environments can challenge an organisation’s governance and control policies.
“Maintaining visibility, monitoring, and compliance across both on-premises and cloud environments is crucial.”
Related:
Becoming “cloud smart” — the path to accelerated digital innovation — Exploring how organisations can become “cloud smart” for accelerated innovation.
