Secure, High-performance Data Storage Fundamental To Effective IT

Sponsored Feature Once again, the world of enterprise IT is bracing itself for a new shock of the new: new application ecosystems, new grades of data resilience, and new data – lots of it, especially of the unstructured variety.

Across multiple vertical industries, IT leaders know their strategies must be adaptive to impending change in dynamics that will push their resources to the limits. Which also means they have to look for insight and guidance from technology providers whose experience reaches across critical technologies.

As the importance of data and data management gain greater criticality with the rise of data-intensive applications, vendors with an already powerful presence in the storage solutions market are especially worth heeding. So global ICT infrastructure provider Huawei Technologies looks well placed to cast perspective on the changes that will shape storage trends.

At its recent 'New Apps ∙ New Data ∙ New Resilience' Innovative Data Infrastructure Forum, the company explored the future, a future evolving toward what it calls the 'yottabyte era' (a yottabyte = a quadrillion GB). Speakers scoped challenges that ranged from the demands of parallel processing of diversified data posed by Big Data and AI, to how collaboration models between data storage and data applications are being reconstructed for emergent data paradigms.

At the same time, cloud-native applications are becoming more prevalent in both private and public data centres. To address this, high-performance container storage is likely to be required.

The IDI Forum outlined how, in 2023 and beyond, the needs of software and data now define IT infrastructure performance and resiliency, explains Alex Zhang, Vice President of IT Product Line at Huawei Technologies

"Not so long ago, only the IT team was concerned about the IT solutions infrastructure – that has changed fundamentally," Zhang says. "Today, the whole of an organisation cares about IT. And that drives some different questions concerning new applications, new data and new expectations for IT resiliency. The IT infrastructure needs to support operational efficiency, business decision-making capability, and R&D capability, for the 'consumers' of IT within the enterprise, as they now are."

Enterprises are moving their application focus towards putting single big systems like ERP or databases into multiple microservices which have to be more agile to keep pace with market dynamics, adds Zhang. Furthermore, most of these applications have to be deployed across multiple clouds (or multiple cloud types), the Internet, and mobile platforms: "They have to be more personalised – and innately socially interactive."

On top of that, the data itself is changing, Zhangsays. "The structure data of those big traditional applications no longer predominates. Huawei has predicted that by 2025 – less than two years away – global volumes of generated data will reach 180ZB, of which at least 80 percent will beunstructured data. So, images, multimedia, with data-intensive backend support systems, including AI. In what seems like a comparatively short timeframe, organisations have been impelled into a situation where they have to be able to adapt to this fully data-centric operating environment."

Business growth drives IT data centricity

Previously it was only IT professionals who had responsibility for IT. But now Huawei is seeing stakeholders who are senior executives approach IT deployment with a different perspective, says the company. "They want to use cutting edge technologies," explains Zhang. "And they are asking 'who is the best IT partner to provision them?' – questions that non-technologists did not previously ask."

This is likely being driven by their different appreciation of the value of data: "They see data as a fluid, highly valuable asset, rather than a static entity that sits in siloes," says Zhang. "And therefore they also recognise that it must be secured against risks such as ransomware."

These changes are occurring against a backdrop of escalated business expansion. Many companies' growth plans include ambitions to extend their operations into new verticals or sub-verticals.

Being prepared to enter a new sector goes to the heart of the fundamental shifts in business dynamics speakers explored at the IDI Forum, Zhang says: "In the past, an organisation that was adopting a multi-sector business expansion model might repurpose an existing customer-facing application for the new market, and that might've worked well enough. But that approach is no longer feasible, especially if a company intends to make optimal use of a data-intensive AI or analytics tool to inform their to-market strategy."

To that end applications must now be designed not just for specific business use-cases, but also for specific types of customer or partner within that new sector. Data used for training AI has to be very specific and high quality for the training to succeed. Storage solutions also have to be able to support multiple data formats – back to unstructured data again – that work in one market, but maybe not so well in others.

Zhang continues: "For instance, Huawei's OceanStor Pacific Scale-Out storage system can be configured to handle AI, Big Data and HPC workloads on a single high-performance platform that supports multiple storage protocols – NFS, SMB, POSIX, MPI-IO, HDFS, and S3 – and that can work within the application layer for some use-cases."

Furthermore, Huawei has been optimising its OceanStor Pacific platform for use in demanding vertical sector applications, such as medical imaging.

"Based on OceanStor Pacific's distributed storage and lossless compression algorithms, Huawei's Digital Pathology solution speeds-up the storage and query of pathological data, and reduces storage space required by 30 percent," says Zhang.  

Challenges of multi cloud and multi data

The consensus cliché among some industry watchers is the that 'multicloud is the new normal', but there is a truism informing this phrase. Increasingly, cloud is utilised as an IT tool as well as a computing environment.

"Enterprises now want dedicated resources to support their business development, whether that be R&D capabilities or production delivery, so they want dedicated high-performance resources. They want a lot," Zhang says. "That's understandable. But there's a problem, and it comes back to data. The reality is that most organisations are still collections of operational departments – executive function, finance, R&D, production, sales, maybe with customer-facing services channelled through the internet. And their respective applications may need to share common datasets, and they end up with multiple copies of data, synced so that it is continuously updated. While this works, it is not really efficient and can be complex from a data management perspective."

But complexity is not an inevitability. Huawei wants to support intelligent sharing and backup between hybrid multiple clouds. "Within one cloud or across data centres we would propose a global file system – a new type of data fabric," explains Zhang. "The data has a local view. So data stored in one data centre is fully visible to other data centres. We don't need to have multiple copies of production data stored across multiple data centres. It's unnecessary and costly. Unified data can be viewed and accessed without being copied and stored everywhere."

Enterprises are starting to realise that while cloud services have always come with a price tag, they have matured to the point where public cloud can be a costlier option than running things on-premises. As more and more data is involved the costs naturally go up. "But that's not the only issue," according to Zhang. "There is the issue of cloud lock-in. If a cloud customer wants to take their data off of a cloud – and there may be compelling reasons to do that – it could cost them a lot of money."

As intellectual property gains in value, organisations may feel less comfortable about keeping it in public clouds, for security reasons both inside and outside of those clouds: "There is also the question of the security exposure of sensitive data in public clouds, or data travelling to and from, and between, public clouds," adds Zhang. "Plus, of course, regulation around cloud governance is being tightened-up, and regulators mandate that certain types of data must be stored on-premises."

In-storage data security

Threats to data resilience threats have evolved from physical damage like fire, flood and seismic disturbance to include malware and other cyber risks. Hence switching from reactive response to proactive defence is necessary to improve storage protection, Zhang explains: "Data storage is becoming the redoubtable 'last line of defence' for data resilience, with more features being integrated into Huawei's data storage products, including ransomware detection, data encryption, secure snapshots, and data recovery in Air Gap."

Huawei's OceanCyber Data Security Appliance serves as a security engine for data storage, offering capabilities that include security policy configuration management, detection, analysis, and defence for various types of storage devices. This platform also builds resilience with ransomware protection features that are customised to the data destination.

The notion of designing data security features directly into storage platforms is not unique to Huawei, but it is indicative of the company's intention to raise awareness of the potential to improve levels of data protection across the entire lifecycle of its customers' data assets.

"At the same time, we make it possible for customers to consolidate their data storage, both in terms of not maintaining unnecessary copies of data – which has a cost, of course – to make maximum use of available capacity, but also simplified management and visibility," says Zhang. "And the tools that enable all those thing have to be easy to use so that effective data storage management is not regarded as in any way a secondary consideration in enterprise IT priorities."

Sponsored by Huawei.