HNHacker News
TopNewBestAskShowJobs

alexkon

732 karma · joined February 18, 2008

submissionscomments
alexkon··on Dremel – Google's tool for analyzing trillion-row tables in seconds [pdf]
Quick view: http://docs.google.com/viewer?url=http%3A%2F%2Fsergey.melnix...

Abstract:

Dremel is a scalable, interactive ad-hoc query system for analysis of read-only nested data. By combining multi-level execution trees and columnar data layout, it is capable of running aggregation queries over trillion-row tables in seconds. The system scales to thousands of CPUs and petabytes of data, and has thousands of users at Google. In this paper, we describe the architecture and implementation of Dremel, and explain how it complements MapReduce-based computing. We present a novel columnar storage representation for nested records and discuss experiments on few-thousand node instances of the system.

Overview by Greg Linden: http://glinden.blogspot.com/2010/12/papers-on-specialized-da...

alexkon··on Unicode 6.0.0 standard published
See the links in the navigation bar on the left for the actual text of the standard.

These are prepublication drafts of chapters 1–7. They may still undergo minor editorial changes. The rest of the text is going to be available Real Soon Now once editorial review is completed. The final publication is scheduled for February 2011.

alexkon··on Exploiting the linker (library loading attacks)
Quick view: https://docs.google.com/viewer?url=http%3A%2F%2Fwww.nth-dime...
alexkon··on Paxos Made Simple – concise explanation of the distributed agreement algorithm
Quick view: https://docs.google.com/viewer?url=http%3A%2F%2Fresearch.mic...
alexkon··on PIRE – really fast single-pass regex matching
Here are the results of some sample benchmarks. These are the same as those used in the paper "Regular Expression Matching in the Wild" by Russ Cox [1].

  500 MB file, regex: .*$
  pcre     31.67 MB/s
  re2     242.28 MB/s
  pire    756.32 MB/s
  
  500 MB file, regex: ABCDEFGHIJKLMNOPQRSTUVWXYZ$
  pcre    153.67 MB/s
  re2     653.76 MB/s
  pire    755.98 MB/s
  
  2 MB file, regex: (\d{3}-|\(\d{3}\)\s+)(\d{3}-\d{4})$
  re2     423,76 MB/s
  pire    775,89 MB/s
Source: Yandex, the company where PIRE originated. [2] (Russian)

[1] http://swtch.com/~rsc/regexp/regexp3.html

[2] http://clubs.ya.ru/company/replies.xml?item_no=30753

alexkon··on Unicode Transformations & Security Vulnerabilities
Quick view: https://docs.google.com/viewer?url=http%3A%2F%2Fwww.casaba.c...

An Internationalization and Unicode Conference presentation by Chris Weber discusses the security implications of various text transformations in Unicode.

alexkon··on Critical Django security updates released (0day, exploit in the wild)
This reminds me of a presentation on regular expression DoS attacks:

http://www.checkmarx.com/Upload/Documents/PDF/Checkmarx_OWAS... [pdf]

Quick view: https://docs.google.com/gview?url=http%3A%2F%2Fwww.checkmarx...

P. S. Corrected wording.

alexkon··on DRAM errors vastly more frequent than previously thought
Did you check whether the downloads had been truncated? Downloads finishing prematurely is a frequent cause for a checksum mismatch.
alexkon··on DRAM Errors in the Wild: a large-scale field study [pdf]
See also a 2007 paper "Failure Trends in a Large Disk Drive Population" by two of the same authors:

http://labs.google.com/papers/disk_failures.pdf

Quick view: https://docs.google.com/gview?url=http%3A%2F%2Flabs.google.c...

alexkon··on DRAM Errors in the Wild: a large-scale field study [pdf]
By Bianca Schroeder (University of Toronto), Eduardo Pinheiro (Google), Wolf-Dietrich Weber (Google).

Abstract:

Errors in dynamic random access memory (DRAM) are a common form of hardware failure in modern compute clusters. Failures are costly both in terms of hardware replacement costs and service disruption. While a large body of work exists on DRAM in laboratory conditions, little has been reported on real DRAM failures in large production clusters. In this paper, we analyze measurements of memory errors in a large fleet of commodity servers over a period of 2.5 years. The collected data covers multiple vendors, DRAM capacities and technologies, and comprises many millions of DIMM days.

The goal of this paper is to answer questions such as the following: How common are memory errors in practice? What are their statistical properties? How are they affected by external factors, such as temperature and utilization, and by chip-specific factors, such as chip density, memory technology and DIMM age?

We find that DRAM error behavior in the field differs in many key aspects from commonly held assumptions. For example, we observe DRAM error rates that are orders of magnitude higher than previously reported, with 25,000 to 70,000 errors per billion device hours per Mbit and more than 8% of DIMMs affected by errors per year. We provide strong evidence that memory errors are dominated by hard errors, rather than soft errors, which previous work suspects to be the dominant error mode. We find that temperature, known to strongly impact DIMM error rates in lab conditions, has a surprisingly small effect on error behavior in the field, when taking all other factors into account. Finally, unlike commonly feared, we don’t observe any indication that newer generations of DIMMs have worse error behavior.

alexkon··on Null-prefix certificate for paypal
If you're using Firefox, set security.OCSP.require to true. There is also UI for it under Preferences -> Advanced -> Encryption -> Validation -> When an OCSP connection fails, treat the certificate as invalid.

I don't know whether it protects against all OCSP attacks but it's obviously better than the default option.

alexkon··on Intel 64 and IA-32 Architectures Software Developer’s Manuals
Intel 64 is often referred to as amd64. IA-32 means i386.
alexkon··on Mozilla proposes an HTTP header to force HTTPS-only connections
There is an improvement over an HTTP redirect to the HTTPS site. Such redirects usually occur in the beginning of every session (or, at least, of some sessions). An active attacker could intercept the redirect at any time a-la Moxie Marlinspike’s sslstrip: http://www.thoughtcrime.org/software/sslstrip/ . But once the browser has received the X-Force-TLS header, it will avoid accessing the site over HTTP at all thus preventing that kind of attack. In other words, the browser will now automatically perform the redirect without issuing any insecure HTTP requests to the site.
alexkon··on Mozilla proposes an HTTP header to force HTTPS-only connections
Contrary to popular belief, it doesn’t make much sense to protect the password without protecting the whole session (that is, without serving most of the site over HTTPS). If just the login page is encrypted, any attacker who is now prevented from stealing your password can still obtain your session cookie that travels over plain HTTP.

A compromise of a single session is bad enough for many services. Set up a filter to forward your email to the attacker’s address? Access private information? Post messages on your behalf? Change your user profile, including your email and shipping address? Usually such actions do not require password authentication when the client already has a valid session cookie. A man-in-the-middle attacker could also intercept your attempt to log out and keep using your session for as long as possible.

alexkon··on Mozilla proposes an HTTP header to force HTTPS-only connections
Here’s the announcement from the Mozilla Security Blog: http://blog.mozilla.com/security/2009/07/27/locking-up-the-v...

I wonder how the initial request could be made secure. With the current proposal, if you happen to work over a compromised network connection the first time you open a site, the site cannot force your browser to connect over HTTPS and the attacker wins.

Daniel Veditz commented on this concern:

“...[W]elcome to the bootstrapping problem. We’re making the assumption that the sites people most want to protect from eavesdropping and tampering are those for which they’ve set up accounts. That means they connected correctly at least once, and this header could be set at that time. In addition, careful users could always add ‘https://’ themselves the first time and avoid that initial redirect.” (http://blog.mozilla.com/security/2009/07/27/locking-up-the-v...)

alexkon··on Google Maps Street View capture... from a bike
What data are they collecting?
← PreviousPage 2 of 2