Hybrid-cloud: Getting the Best of Both Worlds

Having access to, and the ability to generate insights from, high-quality data is a game-changer for organisations. Not only are the majority of enterprises beginning to recognise how data insights enable innovation and optimisation, they are waking up to the need to secure a foot-hold in the expanding digital and data economy. With Cisco estimating that 94 per cent of workloads will be hosted in cloud data centres by 2021, enterprises ignoring the cloud and the data insights it affords will be firmly in the minority.

The public cloud has been generating buzz for the best part of a decade now, yet it is only in recent years that Fortune 500 companies have begun to execute their own cloud strategy. A large component of the debate surrounding whether enterprises should embrace cloud migration is that it is often presented as an either/or proposition. This is to say that enterprises are often presented with the choice of either retaining their on-premises data centres or migrating it all to the cloud. For enterprises torn between choosing an exclusively cloud-based strategy or an entirely on-premise one, both options have their own associated advantages. Staying on-premises offers greater control of data and security. On the other hand, moving to the cloud could lower costs (but only if well-managed) and allow for greater flexibility.

However, what the last couple years have taught us is that the choice between on-premise and cloud is no longer binary. Instead, organisations can get the best of both worlds by partially leveraging a cloud solution while still making use of their on-premise environments. This is known as a hybrid-cloud strategy. The appeal of a hybrid-cloud strategy is strong. It allows enterprises security and control where it is needed while still offering flexibility and lower costs as well. This is precisely why an increasing number of organisations are adopting the strategy. According to the RightScale 2019 State of the Cloud Report from Flexera, enterprises with a hybrid-cloud strategy grew to 58 per cent in 2019 from 51 per cent in 2018. While this trend is certainly good news for cloud providers, it is pertinent to ask what it means for the growth of big data.

ELASTICITY IS KEY

Cloud environments are simply more suited to the requirements and demands of big data than any traditional on-premise infrastructure. This is primarily because the cloud can offer the elasticity needed by most big data deployments. In the majority of data deployments, it is common to see sudden and substantial spikes in resource requirements. For instance, an e-commerce company will typically see consistent, stable levels of traffic throughout the year. However, certain days or events such as Black Friday or Cyber Monday can generate enormous increases in traffic that strain on-premise infrastructure beyond capacity. For e-commerce enterprises whose deployments are integral in analysing the behaviour of users on their site and providing real-time recommendations, having their big data insights fail isn’t an option. To keep providing data insights – and indeed, service – during these traffic spikes, additional servers would be necessary. However, these servers wouldn’t see any use for the rest of the year where their traffic remains at a stable, but far lower volume. This is where the elasticity of the cloud represents a fundamental advantage. In a cloud environment, the same e-commerce enterprise could instantly scale their infrastructure to manage the extra traffic and scale it back down again when it passes. This is both cheaper and easier – when well configured.

Even on a much smaller time-scale, elasticity is still a big draw for enterprises. For example, most apps and websites will have traffic spikes at particular times of the day and lulls at others. According to Nielsen, Tinder consistently sees an increase in active users between 8 pm and 9 pm. Tinder, of course, leverages big data to provide its users with the most likely matches. Maintaining the infrastructure necessary to facilitate these data insights on a 24/7 basis would generate a lot of unnecessary cost for Tinder. Instead, in a cloud environment, scaling up cloud deployments during these peak hours and scaling back afterwards is as easy as pressing a button. By contrast, in an on-premise environment this would involve maintaining physical infrastructure that would only see use in a very limited number of hours in the day.

CLOUD IS EXPANDING BIG DATA

The cloud is no longer a curiosity or buzzword circulated by enterprises. Instead it is becoming an integral part of how organisations adopt big data through simplifying the process and reducing costs. That being said, this is not to say that on-premise environments are going to disappear. There will always be a demand for the tighter control and security offered by physical infrastructure. Regardless, it is inevitable that the cloud will be leading the way for big data going forward.