APAC expert guide to storage management

The rapid pace of digital transformation amid the pandemic has exacerbated the challenges of storage management for many enterprises across the Asia-Pacific region. Many have had to revisit their storage management strategies to keep pace with the rise in remote work, increasingly sophisticated cyber attacks and multicloud adoption.

Take multicloud adoption, for example. As more firms use a variety of cloud services, whether it is to avoid supplier lock-in or to take advantage of the best-in-class capabilities of different suppliers, they will have to grapple with data silos that could dampen their agility in deploying applications.

Kris Day, vice-president for storage, platform and solutions at Dell Technologies Asia-Pacific and Japan, says data gravity will impact a firm’s ability to manage those sorts of complexity. “The larger the amount of data, the more applications, services and other data will be attracted to it,” he says.

Another trend that is shaping storage management is the growing number of edge computing applications that require low latency and generate large volumes of data. Day says organisations will need to evaluate how to integrate data and applications at the edge within their storage management strategy.

Enterprise applications are increasingly being containerised as well. With container adoption growing rapidly and Kubernetes becoming the de facto container orchestration tool, many organisations are looking to simplify their operations, specifically those related to stateful applications, says Sanjay Rohatgi, senior vice-president and general manager of NetApp Asia-Pacific.

“There is a growing need to connect storage infrastructures with Kubernetes clusters and provide data lifecycle management capabilities to drive business productivity and accelerate development cycles,” he adds.

“There is a growing need to connect storage infrastructures with Kubernetes clusters and provide data lifecycle management capabilities to drive business productivity and accelerate development cycles”

Andy Ng, vice-president and managing director for Asia South and Pacific region at Veritas Technologies, points out the impact of cyber attacks on storage infrastructure, which will need to be resilient against ransomware, detect possible breach activities and restore data that has been lost or stolen quickly.

Against this backdrop of increasingly complex IT environments and the unprecedented growth in the amount of data generated by business activity, it has become more critical to approach storage in a strategic way.

As Pratyush Khare, vice-president of systems engineering at Pure Storage in Asia-Pacific and Japan, puts it, “store it and forget it’ is no longer a sufficient mantra for organisations seeking to extract business value from their data.

What’s in a storage management strategy?

A sound storage management strategy contains several key elements. For one, organisations will need to deploy proper classification tools to understand what data they have, where it is located and whether it is valuable or not, according to Ng.

Amid the data chaos, it is also critical to set a retention policy to determine how long data should be retained for operational and compliance needs, who has access to the data and the process of disposal upon expiry. This will help to reduce storage costs while meeting compliance regulations.

Unexpected downtime and data loss are real threats to organisations, especially with the surge in ransomware attacks in recent times. Organisations should factor in high-availability and disaster recovery solutions to minimise disruptions should disaster strike, regardless of where that data is located.

An effective storage management strategy also allows organisations to scale storage capacity up or down accordingly to meet the demands of digital transformation initiatives.

“Organisations can consider incorporating technologies such as hyper-converged and scale-out deployment solutions to eliminate the need to forecast capacity needs and provision in anticipation of future growth,” says Ng.

Joseph Yang, general manager for storage sales at Hewlett Packard Enterprise (HPE), says a firm’s data storage management strategy should also factor in remote, cloud-resident or distributed workloads, along with the ability to support emerging technologies such as artificial intelligence (AI) and machine learning.

“It should address the overarching challenge of enterprise data storage – building storage systems that can support the entire data footprint and enable intelligent and efficient data management at scale to deliver timely insights and business value,” he adds.

Storage management technologies

Besides setting data retention policies, among other strategic considerations, storage management involves a slew of operational tasks, such as resource provisioning, process automation, load balancing, capacity planning, as well as replication, compression, deduplication, snapshotting and cloning, among others.

Dell Technologies’ Day advises firms to consider platforms that provide automation, predictive analytics and customisable automation workflows for all aspects of resource provisioning, load balancing, performance and capacity management, and advanced data services such as data reduction and replica management.

IT teams should also ensure storage array subsystems have the ability to manage tasks such as assessing workloads to apply data reduction without a performance penalty, or management and creation of data replicas in a space- and performance-efficient manner.

Finally, to maximise productivity, Day says businesses should think of comprehensiveness and simplicity while deploying automation workflows. This will deliver hands-off and self-service outcomes for day-to-day provisioning and management tasks.

That is where AI can help, specifically AIOps. “Modern storage management needs to be proactive and predictive to forecast application and storage requirements and AIOps can proactively resolve any issues before they occur,” says Pure’s Khare. 

“AIOps can also benchmark your storage against other enterprises and proactively suggest features and capabilities to turn on to address potential challenges that may become real in future,” he adds.

The cloud factor

With many organisations adopting a multicloud or hybrid cloud strategy and data stored in multiple locations, cloud storage management has become a high priority.

“Modern storage management needs to be proactive and predictive to forecast application and storage requirements and AIOps can proactively resolve any issues before they occur”

HPE’s Yang says having an all-in-one cloud platform that pulls insights on data storage activity will help to drive performance analytics and this is where cloud storage management fits in.

“Cloud provides users with a scalable, top-down view of their full infrastructure, and helps them make better decisions when it comes to shifting data around. Moreover, it creates a unified cloud experience that is as easy to use as the public cloud. It removes the hassle of installing a full suite of storage management software, and users only need to connect their data storage up to the cloud,” he adds.

To alleviate privacy, regulatory and data remanence (residual data) concerns around storing sensitive data on the cloud, Day suggests adopting virtual machine-level encryption, which offers an infrastructure-agnostic approach that is portable across private and public clouds while allowing virtual machines to remain secure during and after the migration process.

Security considerations

A key consideration in storage security is having controlled access to storage infrastructure as well as encryption, says HPE’s Yang, noting that those are the bare minimum features that a storage array should have.

NetApp’s Rohatgi stresses the importance of having flexible encryption and key management solutions to secure sensitive data on-premises, in the cloud, and in motion. “In other words, encryption is only truly effective if it operates seamlessly regardless of infrastructure, and it is only truly effective if it works in-flight as well as at rest,” he adds.

When it comes to backups, Yang notes that a common mistake that some enterprises make is to only create backups in a file server that everyone can access, adding that there is a risk of data getting encrypted by malicious actors, even with regular backups.

A solution is to store backup data in air gap storage – sometimes called offline storage – which is not easily accessed from the rest of your existing infrastructure. Data that is no longer needed should also be deleted as part of data lifecycle management to ensure it does not fall into the wrong hands.

With mounting cyber threats and ransomware posing challenges in data protection, Yang says it is crucial for companies to have backup and recovery services in place to secure their data against ransomware, recover from any disruption, and protect their virtual machine workloads across on-premise and hybrid cloud environments.

Future of storage management

Pure’s Khare believes the next generation of storage infrastructure will gravitate towards “infrastructure as code”, giving developers the speed and flexibility to accelerate software development.

“The virtualisation and abstraction of compute and networking over the past two decades has created tremendous flexibility for administrators and developers. However, storage management has lagged behind the rest of IT in rising to this level,” he says. “It’s time to bring the same agility to storage.”

“Our reliance on data will continue to grow and bad actors will continue to try to take advantage of this. Wherever storage has led, hackers have followed, so whatever happens next, data protection is going to be critical”

That means applying the software-defined model to data storage automation, allowing storage and data services to be provisioned, managed and automated across clouds, making hardware invisible across the entire fleet of storage arrays.

Meanwhile, Khare notes that hyper-converged infrastructure still has its place and will continue to grow in specific use cases and applications. “It’s easy to scale but has a major limitation – both storage and compute resources need to always scale together, which may not be a good fit for many organisations,” he adds.

Veritas’s Ng observes that composable infrastructure has started to move into production environments as organisations deploy new technologies to deliver persistent storage in the ephemeral composable stack. 

However, before deploying tools like Kubernetes distributions to production environments, businesses need to have a good grasp on what data held there can be ephemeral and what needs to be persistent, if they want to realise the full benefits of a composable architecture.

While there are predictions about what’s next in storage management, Ng says one thing is certain: “Our reliance on data will continue to grow massively, and bad actors will continue to try to take advantage of this. Wherever storage has led, hackers have followed, and so whatever happens next, data protection is going to be critical.”