dm_mongodb · HackerTrans

dm_mongodb

no profile record

Submissions

Data centers are too reliable

dmerr.tumblr.com

93 pointsby dm_mongodbvor 15 Jahren25 comments

Using high pressures to create a battery

eurekalert.org

1 pointsby dm_mongodbvor 16 Jahren1 comments

Twitter Open-sources the Home of Its Social Graph

gigaom.com

1 pointsby dm_mongodbvor 16 Jahren0 comments

Prediction: Commutative DB Patterns Ahead

dmerr.tumblr.com

3 pointsby dm_mongodbvor 16 Jahren1 comments

Errors in Database Systems, Eventual Consistency, and the CAP Theorem

cacm.acm.org

41 pointsby dm_mongodbvor 16 Jahren28 comments

300,000 requests / second

qos.doubleclick.net

62 pointsby dm_mongodbvor 16 Jahren21 comments

comments

vor 10 Jahrendiscuss

this in no way effects whether one is trusted for payments.

vor 10 Jahrendiscuss

i can imagine reasons why apple would want or need to be able to update secure enclave firmware without the correct pin entered first. #2-3 are highly speculative:

1) fear of a bug. just the right firmware bug and you have 100 million phones lose data, and perhaps bricked too. unlikely, but consider the cost. i would be worried about this if i were in charge of the iphone project and the secure enclave feature were newish. we can imagine pretty good solutions to this one though, with work and time.

2) out of 100+ countries where they sell phones, over time, some will give them a confidential court order saying they must retain this capability. if a foreign court order, they could refuse to comply, but then would have to exit that market -- infeasible if multiple countries. and a different 'version' for just those countries would be noticed over time by security researchers?

3) they might already have an order as such from the U.S., for foreign intelligence purposes. as mentioned a different international version if noticed is a PR disaster for apple. so the easiest way to comply is just do it that way for all phones.

perhaps they push back on the fbi request because that's the one they can talk about, yet it templates the whole issue.

vor 10 Jahrendiscuss

I don't understand. Can someone please clarify for me. They physically have the phone right? Could they not just read the whole flash, try a pin, write back the whole flash, repeat? (i.e. take the phone apart)

(given it is a 5C with no secure enclave)

vor 12 Jahrendiscuss

you are correct, that's not in 2.8. might want to follow this jira: https://jira.mongodb.org/browse/SERVER-11500

that said it will be a big release in other aspects, with pluggable storage engines now, and WiredTiger specifically.

vor 13 Jahrendiscuss

please tell me most people aren't this silly when they interview folks.

have the carpenter build something (of trivial scope) while you watch. somewhat analogous to coding on the whiteboard in the interview.

come to think of it, it might be darn hard to interview carpenters. but if you watch them work for 2 hours you know everything.

vor 14 Jahrendiscuss

SSL support is in the free distribution but you must built it yourself. One reason would be export controls; another is that creates a dependency on the SSL library for those who don't use SSL, which we found awkward (especially if doing all platforms; the subscriber build just does the popular ones).

So it's available; albeit there is an intent to have a subscriber build with some extra features that are heavily enterprise biased in their usefulness.

vor 14 Jahrendiscuss

re #2: looking at the source code I don't see this behavior. Perhaps in a version from the past? Anyway, please advise or point me to a jira...

In src/mongo/db/db.cpp I see:

    void _initAndListen(int listenPort ) {
        ...
        dur::startup(); // <- this i believe preallocs journal files
        ...
        listen(listenPort); // after journal prealloc i think

p.s. this blog post is a very interesting article overall i think, without commenting on all the specifics...

vor 14 Jahrendiscuss

can you create a Jira ticket for #2 if that is happening it should be easy to fix. jira.mongodb.org

vor 14 Jahrendiscuss

it seems to me that the exchange adding a small random delay (say 1ms) on message reception would dampen the benefits of speed and then there would be less work invested in saving a few microseconds here and there. you still have program trading but a little bit of silly dynamics is dampened.

vor 14 Jahrendiscuss

Goal 1 of the financing was funds to accelerate r&d, to hire engineers to do things like the ones you mention above; in addition more r&d just pushing hard to maximize overall product quality, stability, etc. not sure if those exact things will get done and exactly when yet. I'd definitely like to see the solaris build get some attention. dwight/10gen

vor 14 Jahrendiscuss

i think so yes

vor 15 Jahrendiscuss

obviously there will be a new ceo at some point. time passes.

it will not be schmidt.

vor 15 Jahrendiscuss

it's new.

vor 15 Jahrendiscuss

I imagine Knuth's first version of Tex was pretty darn fast even without 'optimizing'.

vor 15 Jahrendiscuss

I've been an entrepreneur a long time and my take is:

1. Not a myth! Apple, Microsoft, Yahoo, Google, Facebook, Foursquare, ... There are counterexamples but certainly plenty of young ones. I'm 40+ now so no bias. :-) 2. Correct, myth. I could teach you to be an entrepreneur if you wanted to be one. 3. Myth 4. Myth. Gilt Groupe a good example. 5. Myth

vor 15 Jahrendiscuss

Seems to me comparison to C++ might be more interesting. Wonder how that would be given there are 'growable arrays' and such. There's better (not perfect) safety given scoped_ptr, string classes, etc.

vor 15 Jahrendiscuss

consolidation of adjacent btree nodes after key deletions when appropriate. arguably should have already done this, but it's there now!

vor 16 Jahrendiscuss

development agility is a goal of the project too, in addition to scale-out. i've seen a lot of happy users with a single server (or two) and that's all they need.

vor 16 Jahrendiscuss

Single server durability is coming, v1.8 in the fall. We chose to prioritize some other things before that. http://jira.mongodb.org/browse/SERVER-980

We've always clearly said don't build a bond trading system with it. Our philosophy is one-size-fits-all is over; use the right tool for the right problem.

Based on how I like to define the term, there is nothing in the NoSQL space that does full "ACID", including complex transactional semantics involving many objects, on many server clusters. That is ok : the perf + scale problem isn't really solvable if you don't give on something.

vor 16 Jahrendiscuss

Perhaps a better solution is to make trains more cell-based. (Think ATM.)

Each "car" holds just 2-4 people. they run on the tracks and switch on and off at the right stops. Very granular.

But still like a train in that you have tracks and power from the tracks. There is no conductor then, so everything would have to be computer controlled.