Very expensive, unless you only need to use it during peak traffic hours. A month runs you over 700$, and I'm renting a dual quad core with 32GB of RAM, plus bandwidth for 450$ right now.
If your load has predictable peaks it may make a lot of sense to spin up 1 quadruple extra large server rather than the equivalent number of large instances, particularly if you're talking about something like a memcached server.
A small-ish company called 478east(www.478east.com). They offer relatively cheap high-end hardware in their LA point of presence, but lack a fancy control panel like SoftLayer.
Can anyone explain the odd size choices? I've never done any cloud stuff, or even much local virtualization, but I would have expected sized on the low side or equal to an even number (32gb, 64gb) rather than a little higher.
I'm curious about how they provide this much memory, will this memory all be on one machine, or does Xen provide a way to somehow give a VM access to memory on multiple physical machines.
Yeah, the whole 34/68GB is all on a single beefy server. There are systems that let you access remote node's memory in a NUMA type system over fast interconnects, but they're _very_ expensive and have horrible latency.
