Chapter 12 Thoughts about Laptops.
“I work on a laptop specifically so I can work in cafes and pretend I’m part of the human world” Jonathan Lethem
Being in the IT industry, I need laptops for my work. I have what seems like the worst luck with laptops. They end up damaged for one reason or another. When my son was younger, he would rip the keys off. That obviously posed some challenges.
I have had problems with total dysfunctionality, disk failure, power failures (laptop dies and does not respond at all when you try to power it up), motherboard failures, and a host of other failures I can’t even remember. Part of it must be that I’m just travelling with them constantly, but I think some of it is just bad luck.
These experiences led me to thinking about how to build a totally redundant laptop. It’s not a new concept. Total redundancy for components for larger servers is commonplace. You have redundant power supplies, fans, and other hardware components. The idea is to extend this type of redundancy to laptops.
What’s the need? My experiences are probably not unique, although an extreme case. And I know every time a laptop breaks down there’s enormous angst and rebuilding time. Especially for those like me in the IT field, or high powered executives, a laptop becomes a mission critical piece of equipment. So we should apply the same high availability rules for them as you would for high end UNIX servers.
Where’s the market? I believe there will be a sizable market for this. I know I would pay extra for the redundancy. For the amount I have spent going through different laptops, I could have paid 2 times the price and it would still have been worth it.
What do I mean by redundancy? I mean the laptop components are all duplicated. With the size of laptops being what they are at the mini end you could stack all the components side by side. For the larger ones, they could be stacked on top of each other. All the key components would be identical. There would be an interconnect between the two to conduct traffic. Anything installed on the first stack would be duplicated onto the hard drive of the second stack.
Whenever an application processes something, its’ passed through both stacks and duplicate writes are done wherever necessary. If a component in the top half fails, it automatically routes to use the component in the second half. Let’s say for example, the memory board failed. Then the hardware would be smart enough to switch over to using the second half’s memory board. Similarly, if the CPU failed, it would make use of the alternate CPU. There are multiple CPUs on a lot of laptops these days, so you would account for that as well. In case of a major failure, like a mother board, the top half would shut down and the bottom half would kick in.
I don’t think any of this is very difficult with today’s technology. Similar schemes are already used for high end servers; with the cluster interconnect doing a lot of health monitoring. Moving this sort of technology to the laptop would have a niche market of power users who rely on their laptops for their livelihood. Or CEOs’ and other folks who just have extremely critical information on their laptops. I don’t think that the cost would be prohibitive. Even if you double the cost, laptops are sufficiently cheap today that a redundant laptop would be economical.
Whenever a backup component kicks in, there would be a console alert that a component has failed. The user can then instigate a full backup of the system and take the laptop in at their convenience for a suitable fix. In the meantime, you have no data loss and no impact to your business. Valuable documents are protected automatically.
With the advent of smart phones, tablets and other competitors, a larger proportion of laptops are being used only by power users that really need data protection and redundancy. This is yet another compelling argument for the development of my super redundant laptop.
Discuss and enjoy!