Comments for Proton Advisors

Comment on Low Latency Networking… Will You Choose the Light Side or the Dark Side? by Fiber Route

Fiber Route — Thu, 09 Feb 2012 15:20:56 +0000

Wow, this is very well written. Thanks for all the great info! Love your site!

Comment on How Fast is Enough? by Bruce

Bruce — Wed, 02 Nov 2011 03:02:07 +0000

very good analysis. on a tangent, your first paragraph triggered my infosec reflexes to contemplate where Reuters would be if falconry had been commonplace in 1851. tying that to the topic at hand, latency based strategies seem especially vulnerable to disruption by outside factors that interfere with technological foundations of communications and processing. could be an interesting competitive tactic.

Comment on Management Strategy in Technology Sectors by Sheila Callahan

Sheila Callahan — Fri, 26 Aug 2011 07:33:05 +0000

Lucky students.

Comment on Management Strategy in Technology Sectors by Robert Scott

Robert Scott — Wed, 24 Aug 2011 16:58:21 +0000

Sounds like an interesting course. Good luck!

Comment on Building a Low Latency Trading Network, or “Why is my latency so high?” by wbcarley

wbcarley — Tue, 04 Jan 2011 16:53:15 +0000

Thanks for the comment.

No, there isn’t any way to have a signal travel faster than light. There has been speculation about faster-than-light signaling using quantum entanglement, but that is theory, a long way (if ever) from reality.

The answer isn’t breaking the law of physics to greatly lower long distance latency though. Engineering is all about building things within constraints (like physics, budget, etc..)

While it is true that latency is a problem for many applications, to me that indicates either applications that are poorly designed, or use of an application in a way that wasn’t intended. If the application needs to operate over a really long distance, the application designer needs to understand latency and take that into consideration in their design. This isn’t new by the way; See “the Fallacies of Distributed Computing” first articulated by Peter Deutsche (then at Sun Micro) in almost 20 years ago.

That doesn’t mean that it isn’t important to lower latency. It is. But you need to understand what you are trying to achieve and the costs of different alternatives. For some applications (e.g. telephony) the latency needs to be low enough to allow people to speak. Below that threshold, improvement is irrelevant. For other applications (such as many financial trading applications) what matters is not an absolute latency number, but having the lowest latency available (i.e. being first is more important than the time it takes to get there.) etc.. And lowering latency has a cost. That cost needs to be considered against the cost of alternatives. If your trading application has tens of milliseconds of overhead, you may want to spend your money on reducing that before you buy the fastest network.

Brennan

Comment on Building a Low Latency Trading Network, or “Why is my latency so high?” by realdreams

realdreams — Mon, 25 Oct 2010 16:28:46 +0000

When dealing with really long distance(like 5000 mi+), the latency is apparently a problem for many applications. Is there a way to allow the signal to travel faster than the light? Even if the refractive index is 1, the round-trip latency won’t be smaller than 133ms to travel to the opposite side on earth(assuming it’s a shortest route between those 2 points). We probably need something to break the law of physics to greatly lower the long distance latency.

Comment on Why is Google getting into the telco business? by wbcarley

wbcarley — Fri, 06 Aug 2010 13:46:11 +0000

Some further data on Google and telecom, this time regarding Android. Bottom line: As Android proliferates, Google captures more search based advertising revenue (which might otherwise got to Apple through iAd, etc.)

http://www.businessinsider.com/android-revenue-2010-8