How to be a data center hero
FYI, this story is more than a year old
Do storage and IT admins share more in common with Clark Kent or Superman? While IT admins may not be leaping tall buildings in a single bound, they do face a daily battle when it comes to latency and performance issues.
The IT admin is the first point of call when it comes to troubleshooting errors, increasing performance and reducing downtime. To be a hero, an IT admin doesn’t need to fly into burning buildings or foil an evil mastermind’s plan for world domination – they have to guarantee performance and reliability, 24/7.
The biggest challenge admins face is that conventional storage is ill-equipped to deal with virtualisation because it is built for physical workloads. In 2015, about 75 per cent of all workloads in modern day data centers were virtualised and this number continues to grow. Many of Tintri’s client sites in ANZ are now at or close to 100 percent virtualised.
According to IDC, platforms will extend down to the data center infrastructure layer and by 2017, over 65 percent of IT organisations will implement hybrid IT environments that leverage in-house and external 3rd party assets housed in off-site data centers.
If companies have gone virtual but storage is still using LUNs and volumes, it will drive IT admins virtually insane. Conventional storage risks acting as kryptonite – crippling any would-be superheroes ability to deliver performance and reliability.
Speed and agility
Just like a superhero, IT admins need to offer guaranteed speed. Heroes don’t show up once the fight is over: they arrive to save the day in the nick of time, and not a moment too late. When it comes to running a data center, the last thing employees want is to be tied down with latency concerns. But managing and provisioning storage built for physical workloads within a virtual environment is time consuming and complex – even for the most well equipped IT hero.
Conventional storage mixes the needs of multiple virtual machines (VMs), which creates contention. But armed with storage for virtual environments, IT admins can provide each VM with its own lane for predictably fast performance, which is exactly what the IT hero (and the citizens of Metropolis) needs.
Without the limitations of traditional, physical first storage, IT admins can operate six times faster. Operating in LUNs and volumes slows the entire process down and can act as a bottleneck when it
comes to data center requests. Architecting, provisioning and maintaining virtual desktops, databases and servers with legacy storage systems requires many hours of administration.
However when you’re drowning in numerous requests, time is very much of the essence. Using traditional SAN or NAS with VDI storage architectures requires multiple datastores for storing base and replica images, OS disk, persistent disk and disposable disk on different storage tiers. This is too much for any IT admin to handle.
Quality of Service
What IT admins need is a tool that allows them to manage those critical VMs easily, while eliminating the guesswork of Quality of Service (QoS) settings. The trouble with conventional QoS is it works at the LUN or volume level. Consequently, when you set a minimum or maximum input output per second (IOPS), you can only do so for an entire LUN or volume and all the dozens of very different VMs inside get the same level of QoS. By now it’s clear that traditional storage is the nemesis of every IT admin battling through performance and latency issues on a daily basis. Just like any enemy, they need to be removed from the system.
VM x-ray vision
To keep tabs on their data center and all its resident virtual machines, the IT admin needs X-ray vision. The trouble is, no one can see what’s going on inside a LUN or volume. You might know that a LUN is suffering from performance problems, but you don’t know which VM inside that LUN is the culprit. Storage built for virtualisation separates every virtual machine into its own lane, so you can see their exact behaviour. That means you can troubleshoot a latency issue in real-time – even across host, network and storage – and even predict the impact of changes to your environment.
Creating order is all any superhero hopes to achieve: removing LUNs and volumes (villains) and operating in a system built for today’s workloads is the first step to a utopian data center. A data center in which admins can provide performance and reliability (while still having time to make a cup of tea) is a dream for most.
Keeping a data center running smoothly is very much the job of an IT hero. The key is to know exactly which villains you are dealing with and to manage every individual VM’s needs with the correct storage.
Article by CK Chan, senior sales director for Asia Pacific at Tintri.