Monday, January 19, 2015

Stuff I'm reading, mid-January edition

Everybody came back from their holiday break with new things for me to read.

  • One year into the Ebola epidemic: a deadly, tenacious and unforgiving virus
    One year after the first Ebola cases started to surface in Guinea, WHO is publishing this series of 14 papers that take an in-depth look at West Africa’s first epidemic of Ebola virus disease.
  • HC26 Full Program
    Hot Chips: A Symposium on High Performance Chips; Sponsored by the IEEE Technical Committee on Microprocessors and Microcomputers in Cooperation with ACM SIGARCH
  • Why does a single integer assignment statement consume all of my CPU?
    Many processor architectures are less forgiving of misaligned data access and raise an alignment exception if you break the rules. When such an exception occurs, the operating system might choose to terminate the application. Or the operating system may choose to emulate the instruction and fix up the misaligned access. The program runs much slower, but at least it still runs. (In Windows, the decision how to respond to the alignment exception depends on whether the process asked for alignment faults to be forgiven. See SEM_NO­ALIGNMENT­FAULT­EXCEPT.)

    It appears that the original program is in the last case: An alignment exception occurred, and the operating system handled it by manually reading the four bytes from m_data[0] through m_data[4] and assembling them into a 32-bit value, then resuming execution of the original program.

    Dispatching the exception, parsing out the faulting instruction, emulating it, then resuming execution. That is all very slow. Probably several thousand instruction cycles. This can easily dwarf the actual computation performed by Calculate­The­Value.

  • GoGo does not need to run "Man in the Middle Attacks" on YouTube
    I use GoGo a lot. I’ve discovered that their system architecture suffers from “bufferbloat” (the same problem that caused Comcast to deploy Sandvine DPI gear to discover and attack bittorrent with “forged TCP” packet attacks, and jump-started the political net neutrality movement by outraging the Internet user community). Why does that matter? Well, if GoGo eliminated bufferbloat, streaming to the airplane would not break others’ connections, but would not work at all, with no effort on Gogo’s part other than fixing the bufferbloat problem. [The reason is simple - solutions to bufferbloat eliminate excess queueing delay in the network, thereby creating "fair" distribution of capacity among flows. That means that email and web surfing would get a larger share than streaming or big FTP's, and would not be disrupted by user attempts to stream YouTube or Netflix. At the same time, YouTube and Netflix connections would get their fair share, which is *not enough* to sustain video rates - though lower-quality video might be acceptable, if those services would recode their video to low-bitrate for this limited rate access].
  • Tintri: We have ZERO interest in adding compute to storage
    Currently a VMware-focused storage supplier using VMware storage abstractions, such as virtual machines (VMs) and virtual disks instead of LUNs and RAID groups, Tintri is expanding out its virtualised server remit to support KVM comparatively soon and Hyper-V around the end of the year.
  • Five recent results in high-performance data paths.
    The sum of these two facts is that we are currently in a period of systems design in which I/O performance is in its ascent: it is becoming proportionally faster relative to computation. This environmental change is demanding that systems researchers and designers reconsider the parameters in how they architect systems. As evidence of this trend, here are five spectacularly interesting papers that have been published at top systems and networking conferences over the past 12 months.
  • The little book about OS development
    This text is a practical guide to writing your own x86 operating system. It is designed to give enough help with the technical details while at the same time not reveal too much with samples and code excerpts. We’ve tried to collect parts of the vast (and often excellent) expanse of material and tutorials available, on the web and otherwise, and add our own insights into the problems we encountered and struggled with.
  • Why Docker, Containers and systemd Drive a Wedge Through the Concept of Linux Distributions
    There’s been an unholy war raging through the Linux world over systemd for some time. Pretty much everything on a system gets touched by what is selected as the first process on a system and how that impacts everything getting started up. People care a lot about this stuff, and the arguments have been passionate. Nevertheless, Mark Shuttleworth conceding defeat on behalf of Ubuntu marked the last major distribution going all in on systemd. Unless forks like Devuan become successful it’s going to be pretty hard to get Linux in a couple of years time without getting systemd as part of it.
  • OVN, Bringing Native Virtual Networking to OVS
    OVN is a new project from the Open vSwitch team to support virtual network abstraction. OVN will put users in control over cloud network resources, by allowing users to connect groups of VMs or containers into private L2 and L3 networks, quickly, programmatically, and without the need to provision VLANs or other physical network resources. OVN will include logical switches and routers, security groups, and L2/L3/L4 ACLs, implemented on top of a tunnel-based (VXLAN, NVGRE, Geneve, STT, IPsec) overlay network.
  • PCC: Performance-oriented Congestion Control
    Performance-oriented Congestion Control (PCC) is a new architecture that achieves consistent high performance even under challenging conditions. PCC senders continuously observe the connection between their actions and empirically experienced performance, enabling them to consistently adopt actions that result in high performance.
  • What Doesn't Seem Like Work?
    If something that seems like work to other people doesn't seem like work to you, that's something you're well suited for. For example, a lot of programmers I know, including me, actually like debugging. It's not something people tend to volunteer; one likes it the way one likes popping zits. But you may have to like debugging to like programming, considering the degree to which programming consists of it.
  • A Non-Programmer's Introduction to Git
    This is all well and good, but how can I (as a non-programmer) use a tool like Git? Here are a couple examples.
  • Do elite software developers exist?
    We should consider the possibility that someone could have 10X talent (whatever that means) without necessarily generating 10X output volume. Maybe the 10X shows up in quality instead of quantity. Maybe this is why elite developers are not paid 10X the average developer.
  • University Of Chicago's New Free Speech Policy Actually Protects Free Speech
    Rather than actually deal with speech issues on a case-by-case basis, universities have instead enacted broadly-written bans on campus speech.

    The University of Chicago, however, isn't jumping on this particular bandwagon. Its new speech policy is more of manifesto than a policy. It's assertive and it's comprehensive -- not in its restrictions, but in its liberties. It's the outgrowth of a study performed by the school and the conclusions it reaches are decidedly contrary to the prevailing collegiate winds.

    The committee behind the report and policy is chaired by Geoffrey Stone, a professor specializing in constitutional law (and member of the administration's intelligence review task force). Stone is a fierce defender of civil liberties, previously having taken Arizona legislators to task for their First Amendment-steamrolling cyberbullying/harassment bill.

  • Linus Torvalds on why he isn’t nice: “I don’t care about you”
    Following his keynote speech at the Conference in Auckland, New Zealand, Torvalds opened a Q&A session by fielding a question from Nebula One developer Matthew Garrett that accused Torvalds of having an abrasive tone in the Linux kernel mailing list. "Some people think I'm nice and are shocked when they find out different," Torvalds said in response (quoted via multiple Twitter accounts of the event). "I'm not a nice person, and I don't care about you. I care about the technology and the kernel—that's what's important to me."
  • But Where Do People Work in This Office?
    I see a lot of awesome stuff, but where is the quiet area where your big brains go to make world-changing software? Oh, jeez.

It looks like the entire month of January may pass with no rain at all. December's rain is just a distant memory now.

It does mean that I get to continue riding my bike.

But I would rather have the rain.

No comments:

Post a Comment