Ten Years Is A Long Time In Computing … And Towards Using The Public Cloud As a Platform to Teach

Ten Years Is A Long Time In Computing … And Towards Using The Public Cloud As a Platform to Teach

In 2012, London hosted the Olympic Games. It doesn’t seem that long ago, but ten years in computing is actually long time. If we agree with Moore’s Law, we have increased our computing power by a factor of 64 (double every 18 months).

Compute Like It Was 2012

And, in 2012, I connected to the Internet with a speed of much less than 1Mbps and often used my university network to connect, but now I can connect with speeds that are up to 1Gbps (x1,000) — and at home. I also often used floppy disks to install and back up servers and operating systems. Now, with a click of a link, my server backs itself up in less than a minute, and where I can go back to any backup I want within minutes.

For servers, when they crashed, I would have to jump in my car and have to fix them over long weekends and evenings. If you have ever seen the effect of a power outage on a data centre, you will know how stressful it is in getting everything back up and running.

Now, within seconds, I can be alerted about a crash, and for it to fix itself. If I wanted to upgrade our servers, I would have to create a whole business case and go out to tender and get someone to install them. Now, I click on a click and upgrade in less than a minute.

2012, and Building Our Own Cloud

And, in 2012, I produced this little presentation on Cloud Computing:

Little did I know about the true impact that the public cloud would have on our lives. I could see it was the future then, but it was costly to teach with, and where students needed a credit card to access it. So, we build our own Cloud (named vSoC) using VMWare vSphere, and which has been so successful.

We have had bumps along the way as it was built with older HDDs, but we learnt along the way, and could create teach-led infrastructures within hours, and provisioned for each student. In fact, our new system (vSoC 2) has finally overcome most of the flaws of our previous system, and with much more resiliance and improved performance. Overall, it and has (and still is) teaching thousands of students in environments that try to mirror real-life systems:

We thus create environments for running cybersecurity tools that just wouldn’t be possible in the public cloud.

The migration from bare metal to public cloud

For my own website (https://asecuritysite.com), ten years ago, it ran on a bare metal server in 1&1. It was becoming increasingly difficult to maintain the infrastructure, and I was continually on the phone with them and trying to fix failing disk drives and so many other things.

So, I made the move to AWS, and it was the best thing I did. It significantly reduced my costs, and where I could easily create snapshots without using costly backup services. It also significantly improved performance levels and reliability. With 1&1, there were often long outages for maintenance and upgrades. And for a lock-in? Well, I ported my Web site before and could build it again somewhere else. For me, most of my code is compiled, and where I can easily build my servers in another cloud.

And AWS as a platform for teaching?

I gave up using AWS for providing infrastructure teaching for the last ten years. But, this week will be the first week back teaching network security using AWS, and I am really looking forward to it. With software coding, I don’t want to teach coding (as that is boring), I use it to teach methods. And the same is true for AWS, as I just want to teach students about scripting and creating and using networked services. My initial lab involves creating Linux and Windows 2022 servers, and setting up their firewalls:

Our students are now enrolled in the AWS Academy, it is a quantum leap from asking them to put in their credit card details and enroll. There was always a worry that students would leave their instances running, and run away with their credit. Also, students also tend to vastly overprovision, and which can be costly. I have lost count of the number of times that a student has asked for eight cores, 32 GiB of memory and a 500 GiB hard disk. That server could probably run a whole Web infrastructure for a company. For teaching, often a single CPU, and 2 GiB of memory is often fine.

And the future?

Many industry sectors, especially those related to the public sector, have grown up with Cap Ex spending, and with little in the way of commitments to future spending. Every year you commit your spending, and then wait for the next year to plan again. But ,for many companies, the idea of buying hardware and software licences for their information infrastructure is now receding fast, and where companies increasingly are investing within the public cloud. It is now software and cyberphysical.

I believe vast new teaching infrastructures can be built in the public Cloud, and where — with a proper architecture — students can be taught coding, cybersecurity, and many other topics. Unfortunately, we have built our education system behind walls, and where the barriers to access are difficult. We have also taken away the control of teachers to create and manage dynamic learning systems, and where their infrastructure is carefully controlled. I think the time has come to care little about the delivery platform, and much more about teaching methods.

Conclusions

The barrier for the access to education has never been lower. So, let’s lower it even more, and allow every person on the planet to have access to this wonderful world of knowledge and advancement.

Fingers crossed for the lab tomorrow. I just hope that it runs well. If you are interested in collaborating with us in our AWS Academy, please get in contact.