Sergey Nivens - Fotolia

Managing data in a multi-cloud world

Asia-Pacific organisations that are hopping onto the multi-cloud bandwagon need to overcome potential data management issues in order to succeed

For years, Southeast Asian logistics company Ninja Van has been relying on a public cloud infrastructure service to power its operational systems that work out delivery routes and assign packages. But as its business has grown, it has had to contend with a growing cloud bill.

“It’s beginning to become more expensive and some people have been talking about cloud jail, where you get locked in to a single cloud provider,” says Shaun Chong, chief technology officer at Ninja Van. “Soon, you may have to pay up to $1m a month as you scale up your business.”

While some companies may choose to move some systems on-premise to manage costs and avoid being tied to a single cloud provider, the reality is that it is hard to walk away completely from the benefits of cloud computing – faster deployment, utility-based pricing that parks IT spending under a company’s operating expenditure, and access to the latest technologies.

To avoid supplier lock-in, more businesses in the Asia-Pacific region are looking at a multi-cloud strategy, where they take their pick from the best cloud applications, platforms and infrastructure offerings to work alongside in-house systems in their datacentres.

“The increasing willingness to adopt multiple software-as-a-service (SaaS) offerings, coupled with the need to distribute risk in their infrastructure strategy by engaging with multiple service providers, is driving the move towards a multi-cloud, hybrid IT environment,” says Nishchal Khorana, director of emerging technologies cloud and datacentre digital transformation at Frost & Sullivan Asia-Pacific.

Daphne Chung, research director for cloud at IDC Asia-Pacific, says organisations are also using various cloud services to optimise IT costs based on specific workloads or projects, such as finding the cheapest way to deal with test and development requirements for a project that may only run for a short period of time.

“There are a number of other reasons, of course, such as backup or disaster recovery,” says Chung, adding that organisations may also end up with multiple cloud services through mergers and acquisitions, or if a cloud service has a specific feature that is not offered by others.

Data migration and portability are very similar to the age-old problem that storage vendors had in migrating data from different storage arrays.
Paul Haverfield, HPE

With multi-cloud expected to be the norm, Khorana says organisations will have to face the complex task of pulling together a seamlessly integrated data management strategy that will integrate disparate cloud services and automate movement of data across their cloud ecosystems.

Several storage and data management software suppliers are already working with enterprises to identify and manage key data management issues, such as ensuring data is always accessible, having a single pane of glass to manage data across cloud services, as well as security and compliance.

“How portable is your data and workloads?” says IDC’s Chung. “What would it take to migrate them if necessary? How accessible is the data by the necessary applications or workloads? Can you get your data back if need be? How will that work? Is there holistic management across the vendors and does it meet integration requirements?”

Chris Lin, senior vice-president at Veritas Asia-Pacific and Japan, says his firm has been making “tremendous progress” in helping organisations manage their data and applications dynamically in a multi-cloud, hybrid environment, including on-premise virtual and physical infrastructure. “Our customers are already restoring applications in the cloud, and moving applications between public clouds,” he says.

On the data side, Lin says it is critical for enterprises to know where their data is held in a multi-cloud environment. In fact, Veritas has developed what it calls an “Information Map” that offers enterprises visibility into their data through the use of metadata, so they can make decisions about what data to archive, delete or keep on public cloud services or in their datacentres, he says.

For now, Veritas’s Information Map does not offer visibility into data housed in cloud applications such as or NetSuite, although Lin says that capability will be added at a later date.

Data migration challenge

But gaining data visibility is only the first step. Paul Haverfield, HPE’s chief technologist for datacentre hybrid cloud technologies, says organisations will also have to grapple with migrating data from one public or private cloud service to another in a multi-cloud setup.

“Data migration and portability are very similar to the age-old problem that storage vendors had in migrating data from different storage arrays,” he says, noting that data migration was a major technical challenge for a long time until storage suppliers developed specialised features to move data between storage platforms.

Although cloud services are built for lock-in, Haverfield predicts that “Microsoft will start to offer services to specifically move Amazon workloads and data to Azure relatively pain-free and with minimal disruption, which was what we had to do as storage vendors”.

Read more about data management

  • Beyond data backup, extracting more metadata to improve business processes and data management will become more important, says a senior Veritas executive.
  • With the expectation that the future will be based around a hybrid IT environment, NetApp is pitching its data fabric vision as the solution for pulling data across private and public environments.
  • Quocirca analyst Bernt Ostergaard explores the future of data management, a field that is being shaped by different technologies, platforms and capabilities.
  • Independent backup and disaster recovery expert Jon Toigo discusses the importance of data management and testing for cloud backup services.

Much of this only became possible in recent years with the rise of software-defined storage (SDS), which decouples storage from a specific hardware platform. The technology is still not well understood in the Asia-Pacific region, but is gaining traction, with many users evaluating or testing it in their IT environments, says IDC.

Gurpal Singh, IDC’s senior market analyst, says SDS systems are characterised by high availability, easy scalability and are well suited to offer a unified interface for storing and managing data across a multi cloud environment.

“With SDS, one can integrate the infrastructure and provide the IT guys with a consolidated view so that it can be easily managed, then automate tasks wherever possible between multi-cloud workloads with an objective to make data easily accessible across geographic locations,” says Singh.

In a report published in February 2017, Scott Sinclair, senior analyst at Enterprise Strategy Group, noted that it is still early days for SDS and multi-cloud data management.

“With a number of technology providers already offering multi-cloud data management solutions, or at least having the right foundational architectures in place, these solutions have already begun to experience traction,” he wrote. “Given the expected demand, this space could easily become the IT segment to watch over the next few years.”

Next Steps

Backup, archiving and DR best practices in multi-cloud

Read more on Master data management (MDM) and integration