Starting out at OCF as a Hardware apprentice, the HPC industry was completely foreign to me as it was never something I’d learnt or read about during my IT courses in school or college. My apprenticeship was in the hardware department which helped ease my nerves as computer components was something I was familiar with and had worked on before.
From my first day I was thrown straight into the job and headed into the engineering room where my manager began to teach me the very basics of my role and how to effectively rack and cable up a small HPC cluster. It is important that the network cables are plugged in exactly where they need to be while ensuring they are kept tidy and accessible. I was given some pointers by my manager and left for an hour to see what I could come up with. Although I managed to plug everything where it needed to be, it was in no sort of way that would be acceptable on a customer site. The cables were everywhere, tangling between each other and not organised in any manner. My manager came back into the room, removed all the cables, gave me further advice and left me to it again. This was the best way to learn these niche skills as I kept improving each time and being left on my own allowed me to combine his advice with my own initiative for the best results. It was important I got familiar with these basics ASAP as we were scheduled an on-site visit for a customer install the following week.
I was presented with some documentation about the customer install which included a Rack and Point to Point (P2P) diagram. These are the most important documents we must create/use in the hardware team as it tells us where equipment is going to be racked and how it will cable to the network switches so everything can communicate across the network. At first glance, I couldn’t make any sense of these documents, it looked like a load of numbers and letters jumbled together with no real meaning, but my manager began to break it down, going step by step until I was finally able to understand what things meant. To test my knowledge, I was given a little quiz at the end to see how much I had really understood which went well and was a good indicator that I was able to remember the information I had listened to.
The next week arrived, and I was on the way to my first customer site. This was my first time seeing a data centre and it looked like I was in a movie set. There were lots of cables and flashing LEDs from all the active systems. I very quickly had to adapt and get familiar with working in data centre environments including the temperature, noise and lack of moisture. I quickly took off my hoodie, put in my ear protection and found myself leaving very often for sips of water as it is very easy to get dehydrated while in there. Having to deal with all of this while working on my first bit of customer kit was a challenge but I managed and was able to complete my first job!
Following this first site visit, I was then tasked with many more jobs over the coming months which got me increasingly familiar with my role and improved both my problem solving and teamwork skills. I’ve quickly learned that in this role within HPC, you can never be fully prepared for a job as none of them are ever the same. There are too many variables meaning you always need to be on your toes looking out for potential trip ups and then being able to come up with efficient and reasonable solutions on the fly. I have become more independent over time too, going on solo jobs or creating documentation for customers, something I would never have imagined doing on my first day.
I have now passed my apprenticeship at OCF and working permanently in the hardware team but I have never stopped learning. Now that I feel more confident in my role, I have started to look into the other departments here at OCF in my free time and expand my knowledge on how the other teams operate and how their work helps a HPC cluster come together. I have worked with the compute team on how they setup the HPC cluster once it has been physically installed, installing virtualization platforms such as Proxmox and using workload managers like Slurm. These were all terms that flew over my head when hearing them in the office but much more of the HPC industry and how we operate here at OCF is becoming clearer to me now.
It is very surprising how little people know of HPC as it truly is one of the most important industries in the world. Many people are using the services provided by HPC and they may not even know it. Whether it’s a researcher, engineer, student or teacher, all of these people will most likely use the services provided by HPC. The popular rise of AI models in the last few years are powered by the physical hardware of HPC, with ChatGPT alone having 700 million users weekly.
My apprenticeship at OCF has given me many skills and qualities to kick start my career in the industry and has also taught me to be much more independent, not just within my work but in my personal life. The opportunities I have had while working here have given me many interesting and unique memories that I will be able to look back on later in life. I am glad I chose the apprenticeship route as I feel this was the best way for me to develop my skills and gain entry into a career.
Calum Outram
HPC Systems Engineer