It’s hard to believe that the all-flash data centre started over 20 years ago. Fibre Channel has been a solid foundation for this and other storage innovations in the data centre since 1995. It began by enabling shared storage and later led to mainstream adoption of server virtualisation. Today, enterprises are looking to deploy flash-based storage and Fibre Channel is the ideal transport to carry this workload.
Six years ago, the concept of an all-flash data centre was laughable. High costs and low capacity were the bane of widespread flash adoption despite the obvious application performance benefits. But time and innovation marched on.
One of my colleagues predicted in 2013 that flash would be as disruptive to the data centre as server virtualisation. It was prophetic at the time, but it has certainly proven out this year. According to Gartner, all-flash array revenue is expected to grow at a compound annual growth rate (CAGR) of 32 percent through to the end of 2020. That’s phenomenal growth, especially compared to the projected 13.6 percent CAGR for hybrid and traditional disk arrays.
This disparity in growth is a clear sign that all-flash arrays are taking over enterprise storage. The business benefits of an all-flash data centre are obvious: incredible application performance; consolidation through de-duplication and compression; a massive footprint reduction and lower power consumption.
It’s obvious from vendor revenue that the market has moved to flash, but even more importantly, by customer spending intentions. That’s great news for the storage industry and even better news for the Fibre Channel industry.
According to Ibid, when enterprises buy all-flash arrays, they connect to Fibre Channel SANs 69 percent of the time. This leads to what I call “the flash effect on Fibre Channel.” When enterprises connect an all-flash array to a SAN, they end up growing their port count 49 percent of the time. It seems that flash adoption is now the compelling event or catalyst for refreshing a legacy SAN or adding new network capacity.
Redefining business value
The speed of change and innovation around flash also reinforces my colleague’s prediction from 2013. In the span of just four years, we’ve seen flash companies created, acquired, thrive and flame out.
As you consider the new capabilities of storage, think about whether your current SAN will be able to unleash its full potential and benefits. Does it have the performance, value and economics to help you modernise your storage infrastructure for the all-flash data centre?
By getting it right, you’ll be able to redefine business value by enabling more agile IT operations and a seamless transition to all-flash data centre.
Where do we go to from here?
Last year, the same colleague predicted that NVMe will be as disruptive to the data centre as all-flash storage. Although we’re still in the early days of NVMe storage and networked NVMe over Fibre Channel, these technologies are already infusing the industry with new energy and innovation.
NVMe (the protocol) is purpose-built for flash memory and will have a dramatic impact on reducing storage latency due to internal parallelism. To put it more bluntly, it’s finally time to say goodbye to SCSI for flash. SCSI was originally developed in the 1980s, and it’s way too long in tooth to support modern storage architectures.
NVMe over Fibre Channel is the future of high-performance networked storage. In a nutshell, it will provide all of the benefits you know and love today, such as always-on availability, flash-ready performance and massive scalability. In addition, you’ll be able to run NVMe and your legacy SCSI-based storage on existing Gen 5 and Gen 6 SANs.
It’s the lowest-risk and best infrastructure for deploying networked NVMe storage. Best of all, it’s already available.
Phillip Coates is a systems engineering manager for Brocade in Australia and New Zealand. He has architected and installed solutions covering WAN connectivity, data centre and campus solutions, carrier based connectivity, various forms of network security, as well as fault diagnosis and analysis.