Just a few thoughts I’d like to share with you.

None of us who isn’t running supercomputers or one heck of a cluster is there yet, anyway. But as RAM continues to drop in price, I can see the day coming.
Come the day we get a terabyte of RAM in our systems, we’ll find a need for it. After all, recall that great prophet Bill Gates who swore 640K of memory would be all we’d need. Real-time high-definition video editing anyone?

That day, by the way, isn’t as far away as you might think. Violin Scalable Memory will be happy to sell you a terabyte-capable memory device to attach to your server. Indeed it would cost you a few - well more than a few - hundred thousand dollars. Still, you can see it coming.

But while I can dream up applications that could use a terabyte of RAM, that leaves aside the wee technical problem of how you access that much memory. Daniel Phillips, a Linux developer, has an idea: the ramback virtual device.

In short, Phillips is using several cache coherency concepts to make using huge amounts of RAM as a virtual drive a practical approach to speeding up I/O-intensive interactions. For example, with enough RAM, you could run, say, an Oracle database running on Oracle or Red Hat Linux, at real-time speeds.

Of course, Phillips’ idea isn’t the same as coming up with an API for applications to address multiple gigabytes of RAM, but still, it is a very useful idea. Indeed, with its potential for vastly increasing database interactions, I can see it being a commercial success and driving the demand for huge memory.

This in turn will bring RAM prices down to normal user prices.

So, what will you do with your first laptop with a terabyte of RAM? You don’t need to worry about it yet. Well, not quite yet, anyway.