mjb

42 ms·

421.
CAP and PACELC: Thinking More Clearly About Consistency(brooker.co.za)
1 points by mjb 12y ago|0 comments
422.
by mjb 12y ago
The Sinfonia paper ( http://www.sosp2007.org/papers/sosp064-aguilera.pdf ) is interesting, and section 4 does a better job of explaining how this 'minitransaction' model works than this blog post does. The part
423.
Two traps in iostat: %util and svctm(brooker.co.za)
3 points by mjb 12y ago|0 comments
424.
by mjb 12y ago
Using the model from the article, about 33% of the ticket price is total fuel cost, and 16.5% ($1300 of $7900) is 'weight marginal' fuel cost. Combining these two gives us 5.4% of ticket cost on a full airliner is weight marginal
425.
The Operations Gradient: Improving Safety in Complex Systems(brooker.co.za)
1 points by mjb 12y ago|0 comments
426.
by mjb 12y ago
I don't think that's true. The ideas of causality, logical clocks and vector clocks have not been abandoned in any sense. In fact, there seems to be significant new interest in eventually consistent systems over the last few years
427.
by mjb 12y ago
If you've got no coordination then distributed systems are much easier. That's a good design goal, for sure, and distributed database designers spend a lot of time thinking about where they actually need coordination. On the other
428.
by mjb 12y ago
I've commented elsewhere in this thread with a list of better articles and papers. If you've only got time to read one, I'd probably read http://research.microsoft.com/apps/pubs/default.aspx?id=1926.
429.
by mjb 12y ago
As others have said, that article is very misleading. Some better ones: * http://research.microsoft.com/apps/pubs/default.aspx?id=1926... - including a very nice way of thinking about CAP and tradeoffs) * http:&#
430.
by mjb 12y ago
> Is there any information about the expected durability of the new EBS SSD volumes compared to the old Magnetic ones? No, EBS durability has not changed with this release. > Also, is it still recommended to pre-warm volumes? or is
431.
by mjb 12y ago
There is a first-touch penalty with EBS volumes. Follow the pre-warming guide and then try your benchmark again: http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ebs-prewa...
432.
by mjb 12y ago
You can't do online migration, but migration via a snapshot should be quick and easy. Snapshots are incremental, so the best way to get low downtime is to create a snapshot, then when that one completes, unmount the old volume and snap
433.
by mjb 12y ago
If you have some time, and are interested in culture, architecture, history or just New York from an interesting perspective, it's worth reading some more of Scouting NY. There's a whole lot of very interesting material there, and
434.
Viewstamped Replication: The Less-Famous Consensus Protocol(brooker.co.za)
54 points by mjb 12y ago|2 comments
435.
Viewstamped Replication: The Less-Famous Consensus Protocol(brooker.co.za)
1 points by mjb 12y ago|0 comments
436.
by mjb 12y ago
This is a very interesting and useful post. I like how Aphyr started with Linearizability, which despite the unfamiliar name is much easier to reason about (for most) than serializability and most other models. This is well worth reading fo
437.
by mjb 12y ago
> I wish there was a little bit more color around what "availability" really means. The definitions used by Gilbert and Lynch in their CAP proof requires availability for updates at all nodes. Other proofs use availability for
438.
by mjb 12y ago
The classic "Leases: an efficient fault-tolerant mechanism for distributed file cache consistency" ( http://portal.acm.org/citation.cfm?id=74870 ) dating back to 1989 is a good read about these kinds of systems. It
439.
by mjb 12y ago
I don't think you can simplify things that far. Shared-nothing is certainly a good architectural goal, and there are many types of distributed systems where it is both highly desirable and achievable. On the other hand, there are many
440.
by mjb 13y ago
Peter Bailis's talk at Ricon West ( https://www.youtube.com/watch?v=_rAdJkAbGls ) is a good introduction to this topic, and is more approachable than the full paper. This is very interesting research, both as a piece of
441.
by mjb 13y ago
Other languages to add to this list would be Dijkstra's Guarded Command Language, and Promela. Promela is especially interesting because of the (nondeterministic) execution semantics, which provide an extremely interesting way to model
442.
The Essential Leslie Lamport(brooker.co.za)
1 points by mjb 13y ago|0 comments
443.
by mjb 13y ago
You may be interested in looking into some of the research that has been done in other areas on Byzantine consensus. Depending on the system model, solutions that are much more efficient than the Bitcoin block chain (but have other tradeoff
444.
by mjb 13y ago
> I thought that Bitcoin's major technical feat was precisely that it showed that you could solve the Byzantine Generals' problem (which is a generalization of the Two Generals' problem). This sort of thing is extremely se
445.
Snark, Chord, and Trust in Algorithms(brooker.co.za)
4 points by mjb 13y ago|0 comments
446.
by mjb 13y ago
Right. For the same reason, the Paxos protocol allows multiple prepare votes from each acceptor. It makes this safe by requiring that each acceptor tell the proposer of the value for the highest ballot number it has accepted, and requires t
447.
by mjb 13y ago
Years ago I was involved in some work on clocks that needed very high performance at low cost. We ended up with OCXOs GPS-disciplined over some of their frequency and atomic clock disciplined (cesium, IIRC) over another part of the range. T
448.
by mjb 13y ago
That depends on what you mean by accurate, and even then they pretty much suck. Clocks are "inaccurate" in a few ways. One of these ways is very short-term: variations in exactly when the clock edges arrive at frequencies near the
449.
by mjb 13y ago
> The answer is more color range per pixel. Doesn't decorrelating the quantization noise (dither) along with appropriate noise shaping make these two equivalent? It's been a long time since I've thought about this in conte
450.
by mjb 13y ago
Provisioned IOPS EBS volumes are SSD based. From the product page ( http://aws.amazon.com/ebs/details/ ): ""Backed by Solid-State Drives (SSDs), Provisioned IOPS volumes support up to 30 IOPS per GB which

More ›