Skip to main content

Flash storage – The cool at the data centre party

A decade is a lifetime in the IT world, and changes can be unrecognisable. For example, 10 years ago, the iPhone hadn’t even been released. The same could be said for flash storage and its journey over the last decade. Flash has been a hot topic for a while now, and the next phase of flash is starting to deliver solid goodness to organisations. 

Earlier this year, 15TB drives hit the market, providing more options to use flash and increase the overall performance of your data centre. At the IT party, flash definitely remains cool, but are we getting the most out of it?

We've all danced at this party

We’ve been here before. In 1965, Gordon Moore, co-founder of Intelâ, predicted that the number of transistors per square inch on integrated circuits would double every year. This is now known as Moore’s Law. The 1990s exemplified this, with newer, faster, and more powerful processers coming out at a consistently rapid pace. 

This evolution was all about speeds and feeds. And with all that processing power, IT best practices flew out the window, as contention and saturation were solved by throwing more hardware at the problem. When virtualisation became mainstream, speed and feed parties evolved into something that more closely resembled refined dinner parties. Consolidation reigned, resources were shared, best practices were maintained, and waste was minimised. 

We’re now seeing flash going through the same cycle, with new and better flash devices coming out. History is repeating itself while best practices are changing.  

Being a good host

No one argues that flash has many benefits, including high IOPs with low, normalised latencies. Regardless of the size of a business, by investing in flash devices, most if not all see immediate gains in application performances. These performance boosts have the ability to revolutionise the business by reducing IO processing time and providing a competitive advantage. 

However, as IT pros, are we making the most of our flash investments?  Are we neglecting our role as host and assuming that, like a charismatic guest, they don’t need our help making new connections? It’s time to show some love to our flash guests and make them feel welcome. We’ve invested in them and we need to ensure that we maximise and optimise that investment. 

Here are some of my best tips to help you get the most out of your flash device deployment:

  • Manage the utilisation, contention, and saturation: Understand your application data path and the accompanying data growth. Know that the bottlenecks will move from IOPs across spindles to capacity contention on the flash arrays. Know their level. Understanding the contention and saturation points and having a clear and complete picture of this will help to maximise your flash investment.
  • Proper capacity planning: It’s easy to continue to add more applications to flash because of the normalised low latencies and high IOPs, but when flash capacity fills up with data, there is a tendency for potential performance to drop off. Understanding and knowing the limitations and capacity thresholds of your arrays will help you plan for long-term success.
  • Understand the continuous service integration and delivery requirements of your application: We all want our applications to run faster, but this isn’t the one and only requirement for applications. How about availability, or scalability, or backup and recovery? Invest some time in understanding how your key applications work, their different needs, and how you can optimise the impact of flash on your data centre. Remember, if you try optimise for all, you end up optimising for none.

With improvements to flash storage continuing to accelerate, businesses are able to scale and become more efficient and effective. By ensuring that we’re taking a disciplined approach to the planning, deployment, monitoring, and managing of flash, pre-empting any party crashers will help optimise the return on your investment. 

Stay flash and party on.

Image Credit: Welcomia / Shutterstock
Kong Yang, Head Geek, Virtualization and Cloud Practice, SolarWinds 

Kong Yang
Kong Yang is a Head Geek™ at SolarWinds with 20+ years IT experience specializing in virtualization and cloud management. He’s a VMware® vExpert, Cisco®Champion, and active contributing practice leader within the virtualization and cloud communities.