Written on: Jan 28, 2021 9:05:25 AM
Written by: Alex Raben
[Data Management, COOLSPIRiT]
Share this post:
Thank you for heading over to our blog. If you're reading this then you most likely live and breathe the technological world that we all live in - which is full of continuous innovation and development.
During conversations between organisations and suppliers today, subjects such as “Cloud”, “Containerisation” and "Cyber Security" often get top billing, but equally important is the continual improvement of data centre infrastructure and all its elements, as these are, after all, the building block foundations of environments from which organisations deliver their services from - increasing products and services for employees and customers.
Looking back at history and considering how we understood options for data storage solutions, we saw a rapid explosion of storage technology in the late 1980s and 1990s; with a raft of new developments, plus new start-up companies coming to the fore. For example, who remembers the new ‘upstarts’ of Auspex and NetApp in the late 90’s who tried to persuade companies to use file serving technology products to host key applications from?
Over past decades, we have seen storage presentation methods and requirements change as new operating systems and applications come to the market, but one clear consistent change has been in disk technology itself, and its development at a rapid rate - this has essentially revolved around the capacities on individual hard disk drives increasing exponentially, to meet the ever-demanding demands of growing data. Today we see HDDs reaching 20TB capacity in a single drive.
Then there are the performance increases we have all experienced, delivered by flash storage...
A storage consideration for many IT teams today is: can we or should we use flash technology within our storage platforms – whether as an all-flash array or as a hybrid platform? Flash storage is not a new technology. It was developed by Toshiba corporation as early as 1984, but it didn't become used within mainstream IT infrastructure until around 2007 (remember Fusio-IO launching?). Originally, prices were always extremely high for flash storage solutions, but in recent years they have significantly tumbled and made it viable for many organisations. From a technological viewpoint we have seen the development and adoption of Single Level Cell (SLC) flash technology give rise to ongoing releases of newer, and more commercially attractive iterations of flash, with Multi Level Cell (MLC), Triple Level Cell (TLC), Quad Level Cell (QLC) all following each other to the market place - each delivering significant performance increases with ever-decreasing costs, and then there is the next iteration to consider - Penta Level Cell (PLC) flash. But what does this mean to us and how can it best serve the requirements of our organisations?
Undoubtedly, the increased I/O performance (often referred to as IOPS) which flash storage offers over traditional disk storage can be a significant benefit for certain applications and platforms, however spinning disk technology still serves very well, especially for certain workloads (such as sequential read or write operations). One key advantage of flash storage is the much-reduced latency it delivers, which can be a key requirement for some transactional applications and database platforms.
If we consider storage systems within data protection solutions, we are all very familiar with the standard approach of “backup to disk”, where we use a solution to take a backup image of our production data and write this to a disk storage device. Most solutions available on the market today offer this capability, and companies behind these solutions strive to differentiate themselves from each other – for example, some with increased performance and others with highly efficient storage reduction.
Today when discussing data protection, there are two specific points which organisations consider for their implementation designs:
On the first point above, we might choose to consider flash technology to assist with the whole subject of recoverability. At one time, flash technology was only considered for tier 0 critical applications and environment needs - essentially due it is relatively high cost. However, as organisations are now realising the reducing economics of flash technology today, and coupled with the business imperative of resuming operations as quickly as possible - this is driving change and innovation in the data protection implementations. And this is precisely where it can make complete sense to consider an element of flash technology within any data protection solution.
Secondly, Continuous Data Protection is another technology that has been around for many years now, but is now consistently being considered in the 'recoverability' question and how quickly organisations can resume operations in the event of a major incident - helping design and deliver their 'Recovery Point Objectives' and 'Recovery Time Objectives'. Continuous Data Protection differs from traditional backup and recovery which typically uses a single point in time to recover blocks of data from, and instead offers a more transactional view of when to recover from.
Whichever approach is used, the use of flash technology within a data protection implementation will certainly help improve operational times. Furthermore, as data volumes continue to increase, this will help organisations keep compliant with SLAs, RPO & RTO targets.
Looking into the future, flash technology will be available in ever-increasing capacities and at price points comparable, if not below spinning disk technology. The advantages this can offer data protection implementations will allow organisations to focus on recoverability and business resumption more than ever before, and not always reliant on production storage features, such as snapshot technology.
We will see a growth in different uses cases around data management which flash technology will enable. The increased performance in both IOPS and reduced latency will underpin developments in areas such as Artificial Intelligence and Machine Learning, and this will likely lead to more parallel uses for backup data. A simple example could be the use of a backup image on flash storage for scanning via a cybersecurity engine to determine if it contains any anomalous code (ie malware), in turn utilising this information to increase cybersecurity within an organisation over standard practices - this process could take seconds and minutes on flash storage, as opposed to hours or days using traditional spinning disk.
We look forward to the future of data protection, and how flash technology will become embedded in implementations - ensuring the fastest recoverability possible...
COOLSPIRiT would like to thank Paul Brunyee of Arcserve, for his invaluable expert thoughts and insight - helping to create this blog.
If you would like to talk simply call our expert team on 01246 454 222 or email firstname.lastname@example.org