Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
justinsb
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
68 ms
·
601.
▲
by
justinsb
15y ago
Doesn't Netflix use Akamai? If so, there's almost no traffic (when measured on Comcast's scale) from Comcast <-> Netflix, as each movie just needs to be downloaded once, and is then cached on a server living inside Comcast's network
602.
▲
by
justinsb
15y ago
I need to apologize: I hate it when people respond with assertions, rather than facts; and I did that myself there. Sorry! So; taking stemming as a concrete example: * With Lucene/Solr, stemming is done with a pluggable class; you can writ
603.
▲
by
justinsb
15y ago
Thanks for pointing that out. What I've found with aws is that it all looks good until you want to do anything beyond the basics. Then you spend a lot of time doing undifferentiated heavy lifting to get around the aws limitations. With solr
604.
▲
by
justinsb
15y ago
Which service should PlatformLayer add next?
605.
▲
PlatformLayer adds Solr-as-a-service, 1 day after AWS CloudSearch launch
(blog.justinsb.com)
11 points
by
justinsb
15y ago
|
5 comments
606.
▲
by
justinsb
15y ago
All modern relational databases use this approach (it's called ARIES), so all offer the same guarantees as Redis.
607.
▲
by
justinsb
15y ago
Care to explain? I believe for Redis, "appendfsync everysec" is the default. The poster's point was that MySQL and Postgres both ship with something like "appendfsync always", and you have to opt-in to the the less safe mode if you want t
608.
▲
by
justinsb
15y ago
That's an assertion, and I'd love to hear the explanation behind it.
609.
▲
The fair Instagram valuation: $3 billion
(blog.justinsb.com)
6 points
by
justinsb
15y ago
|
2 comments
610.
▲
by
justinsb
15y ago
Abstraction layers are nice in principle, but the issue isn't really the APIs, it is the different semantics. To give a concrete example, on Rackspace's (current) IaaS offering disks are local to the machine and persistent; on Amazon's off
611.
▲
by
justinsb
15y ago
If HP were to cancel it for whatever reason, you can go to one of the half-dozen other OpenStack clouds that will be up and running this year. Or you can install it yourself, on your own hardware; it's open source. Now if Wall Street tires
612.
▲
by
justinsb
15y ago
Or half-hosted on HP, half hosted on Rackspace's OpenStack product (when it is released), so that you don't need to code to two different APIs. Or half-hosted on IBM, half-hosted on Internap, half-hosted on Dell public cloud, half-hosted in
613.
▲
by
justinsb
15y ago
Block storage is like a SAN or Amazon's EBS; Object Storage is like a NAS or Amazon's S3 or Rackspace Cloud Files (Swift). So they're not using Swift for block storage, unless they've built some clever stuff on top. OpenStack does include a
614.
▲
by
justinsb
15y ago
So I presume you've signed up? I believe there are lots of people that love dogs but can't have one full time (travel a lot etc); or who would happily take in another dog for a few days to keep their dog company. I imagine the price will r
615.
▲
by
justinsb
15y ago
This is brilliant; both sides of the market exist and are eager to find the other one. Rover's biggest problem could be that people will happily watch dogs for free / even pay to do so. This is an unusual problem for a market maker (are t
616.
▲
by
justinsb
15y ago
Very interesting - thanks! I look forward to seeing more open-source contributions from DotCloud. This does come out of the experience of running FathomDB, so it isn't an ivory-tower project. FathomDB grappled with the fact that tradition
617.
▲
by
justinsb
15y ago
I hadn't seen OnApp before - I'll check it out. Glad to see you went with your own hardware rather than locking your architecture in to AWS - I hope that'll be a big advantage for you long term. DM me on twitter if you're interested in mee
618.
▲
by
justinsb
15y ago
Dotcloud is closed source, but much more mature - they have customers using it in production. PlatformLayer is open source, but I wouldn't run a bank on it - yet! But the big difference is one of scope: PlatformLayer makes it easy/easier t
619.
▲
by
justinsb
15y ago
Thanks. That's definitely the next step - just a matter of finding the time. NodeSocket looks cool - would you mind sharing a bit about how you built it? Would PlatformLayer have helped?
620.
▲
by
justinsb
15y ago
That hit HN fast! I'm the founder of FathomDB (YC W08) & the primary author of PlatformLayer. PlatformLayer lets you run anything as a service: MySQLaaS, RedisAAS, NginxAAS, DnsAAS etc. It's based on the code & know-how from runni
621.
▲
by
justinsb
15y ago
Good deal! Agree that traditional sessions are best avoided, but good to know that megastore is suitable for session data.
622.
▲
by
justinsb
15y ago
Deal. My initial sweeping claims were incorrect. So: How does Google store its sessions?
623.
▲
by
justinsb
15y ago
I see you work at Google. How about advancing the conversation by telling us how Google stores its sessions?
624.
▲
by
justinsb
15y ago
I know about Paxos. I know when to use it, and when not to use it. Seeing as you came up with the example, using Paxos for an "Nth caller wins" is a really bad idea - the Nth caller to your switchboard likely won't be the one selected. Yo
625.
▲
CloudStack joins the Apache Elephant Graveyard
(blog.justinsb.com)
7 points
by
justinsb
15y ago
|
1 comments
626.
▲
by
justinsb
15y ago
I can't say I've ever put something in my cart, not logged in to Amazon for months, gone back to see it still there, and thought "wow - they must use some sort of webscale storage solution - I'm now going to buy the thing I didn't want to b
627.
▲
by
justinsb
15y ago
Thanks for the WAL pointer; it does sound like concurrency is better now. Benchmarking concurrency against MyISAM is a bit like benchmarking reliability against MongoDB though ;-)
628.
▲
by
justinsb
15y ago
SQLite is a great embedded database, but is deliberately not designed for replacing a 'full' database server. It doesn't support highly concurrent usage, it can't be run as a network service, it has a comparatively weak type model. All of
629.
▲
by
justinsb
15y ago
Here's a 2007 post: http://highscalability.com/blog/2007/9/18/amazon-architectur... In short (and filling in some of the gaps with my interpretation), they used "one big Oracle database" up until 2001. The webapps originally talked direc
630.
▲
by
justinsb
15y ago
They may be known for their home-grown databases, but they shouldn't be known for their use of them: Amazon's primary database is Oracle. Dynamo is used for their shopping carts i.e. for storing sessions; Memcache would probably work just
More ›