[erlang-questions] Lets build a stock exchange!

[G Bulmer]

Brian

Okay, so we are pretty much on the same page.

>> My problem is I don't see why *I* would run a stock exchange
>> application, so I don't see why *I* would run Amazon EC2 instances.
>> ...
>
> I don't imagine that many people would want to run a stock exchange or
> auction software themselves.
Oh, I may be interested if I wanted to run a 'real-time' auction for  
my own stuff. Paying for an EC2 instance for a few hours to avoid  
eBay fees sounds interesting if it's easy. eBay and others  
demonstrate that there is lots of content, and liquidity in auctions,  
but who cares.
I will stop flogging the horse now, it is deceased, and going mouldy.

> But, I think that is not the point.  The
> point is that _many_ of us (potential readers of the book) might
> want/need to build systems that have similar properties (distributed,
> fault tolerant, low-latency, etc.) to these systems.
Agreed.

> At that level, I also don't care what the book uses as an example -  
> as long as the
> example demonstrates the relevant ideas.  To me low-latency is really
> important.
I am interested in low-latency too.
In my simple machine-to-machine experiments, I observe approx. 1% of  
message round-trips to be more than an order of magnitude slower than  
the other 99%.

G Bulmer

PS - This is the sort of problem that I would attack using DTrace, if  
I could see the function call or message send/receive within each  
Erlang node.