NVMe – a revolution in storage

NVMe is a new interface standard for storage, designed to optimise the higher performance of newer flash drives.

Non-volatile memory express (NVMe) is a relatively new interface standard for storage, designed specifically to enable the full potential of faster storage technology, such as flash, to be realised.

This article explores how NVMe promises to revolutionise storage as people know it, and how this new standard can impact business, in particular bringing new value to big data analytics and on-premise storage.

What exactly is NVMe?

NVMe is an alternative to existing interface standards for transferring data between a host system and a storage device. Non-volatile memory (NVM), such as flash, is storage medium that does not lose data when the power is cut off and that offers vastly improved performance – 1 millisecond latencies compared to 20 milliseconds with traditional spinning media.

>See also: The journey of data storage: what’s next?

NVMe was developed specifically for use with NVM such as NAND flash technology and has been optimised for next-generation NVM. NVMe further improves the performance of NVM and flash media, meaning it is typically 100 times faster than mechanical disk drives.

Existing interface standards such as SAS, SATA, SCSI, and ATA have been unable to keep up with performance improvements in storage media. This has resulted in bottlenecks for even the highest-performing storage solutions, limiting performance and meaning customers did not get full value out of expensive flash storage solutions or their compute environments.

NVMe demonstrates significant improvements in performance compared to these existing data transfer protocols, which are designed to work with mechanical disk drives.

For example, NVMe devices offer memory-to-memory transfers (bypassing the CPU) at sub-millisecond latencies and can exceed 1,000,000 IOPS, which substantially improves bandwidth and performance.

>See also: Enterprise storage in 2017: trends and challenges

What’s more, NVMe, as with most all-flash and hybrid arrays, can lower Total Cost of Ownership (TCO) through lower power consumption and smaller data centre footprints.

Real-time big data analytics

Real-time big data analytics has the power to transform a wide range of sectors, from retail to government, health to business and even transportation. It enables organisations to process and analyse vast volumes of data, to make informed decisions about fleeting opportunities and to respond quickly to potential problems.

This is of widespread benefit to organisations giving them a competitive edge and highlighting areas to improve, as well as to customers who receive a more personalised, tailored experience.

Supermarkets and loyalty programmes for example can leverage big data analytics to assess shopping habits and offer suitable discounts. The same goes for airlines which assess passenger habits, preferences and search history.

Having access to large volumes of invaluable data such as these enables instant business agility and the ability to make fast decisions.

With sophisticated analytics programmes, organisations are able to analyse these data sets in order to hone their business strategies, tailor their marketing campaigns and improve customer satisfaction, making them more competitive, and ultimately helping them to grow their businesses.

>See also: Is Hadoop’s position as the king of big data storage under threat?

Airlines also apply big data analytics to data generated by their aircrafts and by the flights themselves. It is estimated that a Boeing 787, a highly connected aircraft, creates upwards of half a terabyte of data per flight.

Pilot reports, incident reports, control positions, warning reports and sensors from the engines among others, create vast volumes of data. Airlines are making use of real-time big data analytics in order to monitor and analyse this data to enhance safety.

When it comes to cars, it has been estimated that by 2020 connected car services will account for nearly $40 billion in annual revenue. In addition, it is predicted that self-driving cars will produce 4 TB of data per day.

Remote diagnostics, automatic crash notification, usage-based insurance, traffic management and autonomous driving are not far from becoming a reality – and it is big data and big data analytics that will enable this transformation.

But what does this have to do with storage and NVMe? Sophisticated big data analytics platforms are of no use if they cannot access the required data quickly enough.

Performance increase in storage media and interface standards, as discussed above, means that business agility and real-time data analysis is now an affordable and viable proposition.

>See also: Building a smart storage strategy

Mechanical disk-based storage solutions were never designed to offer the performance necessary to analyse large data sets such as shopping habits or flight data in real-time; however the characteristics of flash storage and NVMe are perfectly matched to such applications.

Flash and NVMe maintain high levels of performance when handling large volumes of data, enabling high-level analytics programmes and platforms to operate to their full potential.

Flash storage has improved the speed and performance of data storage, and NVMe has been designed in order to optimise this, guaranteeing customers see the full value of expensive flash arrays.

By ensuring memory-to-memory transfers at sub-millisecond latencies and avoiding the bottlenecks caused by other interface standards, it is safe to say NVMe’s innovative technology is revolutionising the way we access, use and analyse data.

On-premise

As the need for businesses to store, process and analyse large data sets grows at an unprecedented rate, the decision of how and where to store data is becoming more important than ever.

Until recently, it seemed that the cloud offered levels of flexibility, scalability and cost-effectiveness that were not attainable on-premise. In addition, spinning disks did not offer the performance needed for businesses dealing with big data sets, and for most organisations it was not financially viable to use multiple flash drives.

>See also: Driving research for new drug therapies using big data storage

NVMe is revolutionising the cost and the performance of on-premise storage, meaning that organisations do not have to turn to the cloud in order to achieve their business objectives.

They can store and analyse their data on-premise more cost-effectively as flash and NVMe require less power and offer higher densities, lowering overall data centre costs.

Also, newer, faster technologies are becoming more affordable. For example, a 1 TB internal flash drive is available at less than £300 while just three years ago the same device cost over £2,000.

What this means is that businesses, should they wish too, can more easily take advantage of the benefits of on-premise data storage while meeting requirements in terms of speed, flexibility and costs.

These benefits include control and flexibility over how they run their data centres, allowing the IT department to respond much more quickly and easily to business opportunities, online attacks or system failures for example.

>See also: Flash storage: transforming the storage industry

IT professionals can avoid the constraints and benchmarks of a cloud provider and can choose to increase performance through various flash and NVMe options.

So is NVMe really a revolution in storage? Arguably it is, but even if you disagree, it is certainly causing a significant shift in the storage industry and opening up many possibilities.

 

Sourced from Narayan Venkat, CMO, Tegile Systems

Avatar photo

Nick Ismail

Nick Ismail is a former editor for Information Age (from 2018 to 2022) before moving on to become Global Head of Brand Journalism at HCLTech. He has a particular interest in smart technologies, AI and...