Oleg Zabluda's blog
Monday, October 15, 2012
The orthodox probability theory I was taught in high school, is now being called "Reactionary Frequentism", staying...
The orthodox probability theory I was taught in high school, is now being called "Reactionary Frequentism", staying in the way of fixing Foundations of Quantum Mechanics, discovering Higgs to 5 sigma, and generally making no sense at all. Glad July 2012 issue of Physics Today set me straight with Bayesianism. [1] [2]

"Commentary Quantum mechanics: Fixing the shifty split" by N. David Mermin aka "Frequentism vs Quantum Bayesianism"

"Discovery or fluke: statistics in particle physics" by Louis Lyons aka "Frequentism vs Classic Bayesianism"

Book review: "The Theory That Would Not Die: How Bayes’ Rule Cracked the Enigma Code, Hunted Down Russian Submarines, and Emerged Triumphant from Two Centuries of Controversy" by Scott L. Zeger, Reviewer

[1] When I told all this to Marianna Dizik, as some kind of a revelation,  she said she always intuitively adhered to the progressive forces of Bayesianism, even before it was cool, and, in fact, before she knew about their existence. Actual experience rather then classroom tests, quickly would have set her head straight even if she wasn't built like that from birth.

BTW, in USA, they teach provability theory and statistics in elementary school. Different needs for living under Politburo and Gosplan vs whatever we have here.   

[2] Although I did dabble into Bayesianism here:


Great contemporary composers are constrained by a bottleneck of asking actual musicians to perform their...

Great contemporary composers are constrained by a bottleneck of asking actual musicians to perform their masterpieces for them, or, even worse, having to learn to play piano themselves. Great songwriters, additionally bottlenecked by the need for a singer, because exposing others to their own singing is generally prohibited by the Geneva Convention (except for the terrorists). Humanity is missing out on great creations of art, because computers just don't cut it yet. Fortunately, finally, some great use of your taxpayer dollars:

"Musical rhythms: The science of being slightly off" from Physics Today, July 2012.


Samples, vote, etc:


Some experiments into what can be achieved with flat L2 networks, when 10G network is faster then small number of...
Some experiments into what can be achieved with flat L2 networks, when 10G network is faster then small number of local drives.

Currently standard way to scale datacenter-scale networks is with fat trees (edge-aggregation-core). Their limit is achieved due to the cost and technological availability of core switches. A cheaper alternative is a "flat" Clos network made out of commodity switches [1]. Its biggest advantage is cost and same-speed links throughout at the cost of increased latency.

But single-hop latency, currently is on the order of tens of microseconds, while hard-drive latency is on the order of tens of milliseconds. By picking large enough block size (8 MB in the paper), this can be mitigated, as well as the difference between rack-local hops and global hops. 

Then the win comes from the fact that network bandwidth is on the order of 1 GB/s, while spinning hard drive bandwidth is ~100 MB/s. Locality is still very important for flash (~1GB/s, 25 μs). and RAM (~100 GB/s, 25 ns).

"Flat Datacenter Storage", Microsoft Research, 2012 by
Edmund B. Nightingale, Jeremy Elson, Jinliang Fan, Owen Hofmann, Jon Howell, and Yutaka Suzue

"Towards a next generation data center architecture: scalability
and commoditization", Microsoft Research (2008) by A. Greenberg, P. Lahiri, D. A. Maltz, P. Patel, and S. Sengupta.

"A Scalable, Commodity Data Center Network Architecture",UC San Diego , 2008 by Mohammad Al-Fares, Alexander Loukissas and Amin Vahdat

[1] The UCSD paper infuriatingly keeps referring to their Clos topology as "fat tree", although it contradicts established terminology, is not even a tree, and not fat.


"Sun Datacenter Switch 3456 Architecture White Paper" (has diagrams for T=5  and formulas for arbitrary number of tiers T.

"PortLand: A Scalable Fault-Tolerant Layer 2 Data Center Network Fabric" 
a three-stage fat tree built from k-port switches can support non-blocking communication among k^3/4 end hosts using 5k^2/4 individual k-port switches 

"Scale-Out Networking in the Data Center" 


Powered by Blogger