When hyper-converged infrastructure may be your best bet

Enterprises of all sizes are moving an increasing number of their applications to popular public cloud services. It’s true that for some corporate apps, particularly those supported by a software-as-a-service (SAAS) model, the cloud is a popular option.

But for other applications – especially long-running legacy systems – using hyper-converged infrastructure (HCI), which combines compute, storage, and networking in one appliance, may be a better bet.

Tech execs gathered in Melbourne for a recent Computerworld event sponsored by Lenovo to discuss their application strategies, which systems they have moved to the public cloud, and if they are taking advantage of HCI technologies.

Bernard Wansink, chief information officer at Schiavello, says the construction firm manages extensive infrastructure in-house, and to date, it has moved minimal services to third-party service providers.

“Our current approach is to shift services to the cloud that require continuous updates and on-demand scaling but do not contain any group IP,” says Wansink.

“Hyper-converged infrastructure is currently on our radar – driven by Internet of Things and risk management initiatives. We are building use cases on the business need, which are due for completion in 2018.”

Glenn Goodman, head of business systems and data quality, whole lending operations, at ANZ, says the majority of the bank’s infrastructure is housed internally in data centres with some individual applications being made available to staff from the cloud. These are not typically integrated with internal applications, says Goodman.

“Although our infrastructure manager is well aware of the HCI model, it is not something we are using yet,” says Goodman.

“However, we are trialing and adopting similar in-house technology in parts of our data centres for similar reasons: scalability and efficiency, and shared instead of allocated machines.”

“This type of development would logically lead towards using hyper-converged infrastructure at some stage,” he says.

Public cloud causing ‘bill shock’

Until recently, the public cloud was widely viewed as a possible one-size-fits-all solution for IT. But in a massive shift, many enterprises and early adopters are taking steps to back track as “costs spiral out of control, causing bill shock and mission critical applications to stall and even fail,” says Andrew Silvers, solutions alliance manager, ANZ at Lenovo.

This is a key driver of demand for HCI solutions. The exponential uptake of these technologies is fueled by the clear benefits they deliver to enterprises, particularly cost and time savings, and the scalability and agility the infrastructure provides, says Silvers.

One of the biggest advantages of deploying HCI is the savings in man-hours it brings by simplifying the daily administration of infrastructure, Silvers says.

Hyper-converged solutions are far simpler to manage, monitor and maintain using a single, simple interface as opposed to three separate and potentially disjointed tiers – and as a result, they can easily be managed by an individual, he says.

“Hyper-converged systems also offer substantial opportunities for organisations to scale with simplicity. Nodes can be ordered, racked and brought online quickly compared with a traditional three-tier architecture.”

Using the cloud is a good solution in many instances but it needs to be for the right workload. Instead of putting all the eggs in one basket, a solution for enterprises is to look at hybrid cloud strategies, says Silvers.

“New applications designed for the cloud or applications used less than a third of the time during a three-year period are perfect for public cloud,” he says.

“However, higher use of legacy workloads are generally better suited to in-house infrastructure or private clouds. This is where there is potential bill shock as high use applications can also require frequent data transfers that can be costly and time-intensive, which results in underwhelming performance.”

Silvers adds that in a hybrid cloud environment, it’s also possible to host apps in the cloud but the data outside. This means organisations are in control of their data and can get applications back up and running swiftly following downtime.

Another key consideration when looking at deployment locations is compliance, says Silvers. For most enterprises, there are specific legal obligations, regulations, geographic rules, and even internal mandates such as privacy agreements with end users, which need to be adhered to.

“These requirements can change from one customer to the next so a company needs to be able to adapt its approach as needed. For example, we have a medical imaging customer using hyper-converged infrastructure to deliver the scalability and flexibility cloud provides, as well as consistent performance, cost, and meeting of compliance requirements.”

ANZ’s Goodman adds there are several factors at play when deciding whether or not moving mission critical apps to the cloud is a good move. Firstly, he believes many of the original concerns around cyber security and safety have diminished over the past two years.

“The major players in the cloud market most likely do these types of security controls and monitoring as well as or better than most organisations,” he says.

“This probably rings true for many small organisations that can go ‘all in’ with their applications. Our problem as a larger organisation is that moving parts of all total system stack to the cloud but not the dependent, integrated parts could actually reduce performance and create additional performance risks.”

ANZ also operates in a highly regulated environment with many cross-border considerations, which gives rise to concerns around data sovereignty, he says.

The last factor is whether older type applications can even operate on newer infrastructure provided by cloud vendors. An overhaul of operating software, code, and database structures may be required to bring the provider up to speed, he says.

“I am aware that we are currently trying to more clearly define our rules on how or when cloud-based applications can be used but this is still being worked through,” he says.