justinsb

93 ms·

811.
by justinsb 17y ago
But the users are going to issue those N queries while you're still at the whiteboard talking about big-O notation. Exposing a full query language to the database system means that you have the opportunity to optimize the query fully. Limi
812.
by justinsb 17y ago
You can't rule out queries to suit your implementation; if the data is needed and can't be computed within the constraints, then the user will just run multiple queries to get that data. For example, a key-value store can run simple lookups
813.
by justinsb 17y ago
I agree that reasoned and clear debate is the best approach. I've tried to make the case 'pro' by citing the impact of PageRank/MapReduce/GFS, and suggesting that the impact is no less than reshaping the way we think about computing. So wh
814.
by justinsb 17y ago
I'm arguing simply that Google's papers were IMHO the most influential Computer Science papers of the past decade, and they wouldn't necessarily have been published without patents. The patent system is designed to encourage exactly the so
815.
by justinsb 17y ago
But people do read the papers, and I'd argue that the papers are only published because the patent has been filed. I'm willing to bet that 80% of the hackers here have read the MapReduce paper (and for anyone that hasn't, you should!), but
816.
by justinsb 17y ago
I'm the author... and yes, the 'implication' is exactly what I'm saying. As I said in the post, you can argue that Google would have published in a world without software patents, but the purpose of patents is to incentivize publication wh
817.
by justinsb 17y ago
I agree ... it's a major pain on EC2 to have separate images that are only used for the smallest instance sizes. While 32 bit vs 64 bit pointers do save RAM on these small machines, I think you make a good point in an amusing way, and you
818.
by justinsb 17y ago
Interesting links - thanks. There's definitely a continuum here - the full ACID model at one end; key-value stores at the other. It looks like Google is moving in the right direction - adding more ACID - which certainly plays to my person
819.
by justinsb 17y ago
Completely agree with you on the theoretical level; it's a (lazy) shorthand to contrast the typical NoSQL trade-offs with the ACID model that most relational databases employ. In practice though, I think if you introduce (multi-'row') ACID
820.
by justinsb 17y ago
As Jonathan has pointed out, the article has lots of inaccuracies, but it also has a very good explanation of the NoSQL Faustian pact: "Cassandra doesn't do joins... doesn't guarantee referential integrity, where the user knows the data be
821.
by justinsb 17y ago
I think I was pretty open about the fact that it's obviously going to be our viewpoint that we don't think people should do NoSQL. However, it's also because I genuinely believe SQL is the right decision. There may be use-cases out there
822.
Scoble interview with FathomDB Founder (W08)(building43.com)
17 points by justinsb 17y ago|2 comments
823.
by justinsb 17y ago
Agreed, but I think there's not as much need with linode & rackspace. I believe they pretty much work as you would expect, and they're sufficiently open that if you asked questions my guess is that their default position would be to te
824.
by justinsb 17y ago
That's a great way to look at it. The problem here comes when you want to add a new feature, and you find that your design doesn't allow for the new queries needed. The relational model and data normalization are reasonable at stopping you
825.
by justinsb 17y ago
Yes - that's my understanding of the goal of NoSQL. As for whether they've achieved it, that's a more complex question :-)
826.
by justinsb 17y ago
I believe that simplifying querying is a non-goal for NoSQL, in fact most NoSQL databases actually push more of the burden of querying onto you. CouchDB is arguably a bit better here with its concept of views. SQL is incredibly powerful f
827.
by justinsb 17y ago
Is your app pushing the limits of SQL databases? Is there any reason to look at NoSQL other than the fact that it's 'cool'? Currently all the NoSQL databases are very early adopter products, and each have their own strengths and weaknesse
828.
by justinsb 17y ago
Though perhaps there's a selection bias at work? Maybe 'competing' venture funds are doing more thorough IP due diligence?
829.
by justinsb 17y ago
As someone that ran a company for a number of years that sold solutions for session storage, the best solution is not to store them... Encode your state in cookies instead. Of course, that's much easier said then done, especially given the
830.
by justinsb 17y ago
I believe the metadata file could equally well be generated on the client side, as long as it's stored on the server.
831.
by justinsb 17y ago
Are you volunteering to fix the bug if I can reproduce it? That would be very kind of you :-)
832.
by justinsb 17y ago
Everyone knowledgable about nginx (that I've talked to) says not to use it as a load balancer though. Best practice seems to be to front haproxy with nginx. I've personally encountered some pretty serious issues when a back-end server isn'
833.
by justinsb 17y ago
Awesome ... was that for security or for redundancy? Did you have a solution for what would happen if the IP address changed?
834.
by justinsb 17y ago
I thought DNS was supposed to try backups servers automatically... any DNS experts able to explain what's going on? Some of the ultradns servers are returning (correct) values, others simply not responding.
835.
by justinsb 17y ago
It is DNS. If you put the EC2/S3 address into /etc/hosts, the services work fine. Affecting lots of other big websites as well apparently (target, salesforce) because they all outsource to UltraDNS
836.
by justinsb 17y ago
Right, I agree that scaling out is great. However, bursting is also a great feature; most loads are bursty (you're caching data and have a cache miss or timeout, you request an admin page which does a lot of computation.) All time-sharing
837.
by justinsb 17y ago
I think the idea is to get away from the messy situation you see in TPC database benchmarks, where the environments are highly tweaked by the vendor just for that TPC test, and this bears no relation to the performance you actually get as a
838.
by justinsb 17y ago
Didn't EC2 sell the cloud on the idea of being able to burst? Isn't it better to be able to do that for short periods without having to fire up a new server, configure it, and pay for a whole hour? Your point is very valid from the point o
839.
by justinsb 17y ago
I think he's comparing an EC2 small (we can tell because he says the memory is between 1GB and 2GB); it looks like the headings on the table are wrong. I think he's comparing this to a Rackspace 1GB, which is actually cheaper than EC2's sma
840.
by justinsb 17y ago
Thanks, I was a little off on pricing - I didn't realize Linode started at $20. Rackspace is $11 / month for the 256MB slice, no bandwidth, though there's no bandwidth included in EC2 so that's apples-to-apples. I doubt bandwidth costs wou

More ›