Tech Field Day is a wonderful way to learn the strategic direction of enterprise IT companies, and perhaps no one gets a better view of that than I. After all, I attend Wireless Field Day, Networking Field Day, Storage Field Day, and Virtualization Field Day, hearing about the latest developments in tech, from distributed WiFi designs to software-defined networking and network function virtualization to software-defined storage to cloud. And through it all, there is one unifying thread: Automation.
To be sure, automation is nothing new to IT. In fact, we’ve been fighting the automation battle for decades, with companies and technologies focused on operational efficiency rising and falling along the way. But these were mostly hype. I’ve never before seen the kind of hope that exists today.
Storage administrators have long been lining up on either side of the automation divide. Many cling to the old ways, crafting disk layouts and LUN access in Excel and fighting any changes tooth and nail. But others have accepted that they simply cannot control storage configuration anymore and are embracing technologies like automated storage tiering, automated data placement, and scale-out storage.
You should also read How I Learned to Stop Worrying and Love Storage Automation and Follow the Yellow Brick Road to the Software-Defined Future
Flexibility is the true harbinger of automation in storage: By hand, it’s possible to craft an architecture with fantastic performance from simple ingredients but nearly impossible to make this same environment change and grow as new demands appear. Although generally less elegant and efficient, automated systems have the benefit of continuing to function once needs change, clients are added, and I/O grows.
Growth is the reality of most data centers today, making automated storage a necessity. Administrators have to relax and allow the system to do what it chooses, accepting that it may not match their own designs.
And this is incredibly hard for many to accept: It’s become common at my Truth in IT seminars to get pushback from storage administrators in the audience arguing that the best way to build IT systems is by hand, with carefully crafted RAID sets and data placement. Although they have a point in the likely small, static datacenters they work in, this approach is simply impossible in a dynamic environment.
Stephen’s Stance
The only way to build a datacenter with flexibility and scale is automation. And this is as true for networks and servers as it is for storage. IT architects increasingly design integrated and automated systems, not static interconnects. They must learn scripting and look for solutions that are responsive to changing demand. And they have to start getting excited about companies playing in this space.
Image Credit: Automatic, Watford Gap Service Station, England, 10 March 2008 by Dr John2005
Andy Konecny says
We are always automating things in IT. Once upon a time it was automating the setting of IRQs on add on boards, or choosing where in memory a program would go, or driver installations when a new device was added. We keep inventing new things to do that initially require close attention, but as we figure it out we can automate some or all of it and move on. Some people happen to come on board IT when a particular new technology comes along, they learn it really well, and then are lost when that skill gets automated away.