Your $5 Pentium 3 server isn't the largest retail website on the internet making $61 billion a year.
Having seen a lot of the code that Amazon runs on, and having seen first-hand the scale that it runs on, I'll say this: it's not perfect, but it's remarkably well-engineered, and a hell of a lot better than most snarky HNers could do.
But that's the point. Most people don't need anything that well-engineered. Compared to more traditional hosting solutions from quality providers, AWS has terrible uptime and at a much higher cost for the same amount of resources. Two VPS'es from two different providers in a simple failover configuration with an anycast DNS solution would be simpler, cheaper, and much more reliable.
Wow, apparently that last comment really hit a nerve, as several people decided to downvote it, but not a single person actually refuted any of what I said. I was under the impression that downvotes were more to be used against trolling or flamebaiting, and not just opinions that people disagreed with. Considering everything I said is quite easy to verify as being true, this downvoting just strikes me as kind of intellectually dishonest. I expected better from HN.
Yeah, I think there's a misunderstanding somewhere. Some people think I believe a $5 computer could handle amazon.com's traffic, which is clearly preposterous.
I know that almost all of my downtime comes from when I overengineer things. And I don't need to "patch my kernel" because my OS doesn't have kernel holes once a week. Linux isn't the only Unix OS out there.
Today, a lot of sysadmins believe that "LAMP" is a synonym for webserver, and consequently there are a bunch of webservers serving static content on a machine with way too many moving parts. Complexity is bad.
"Things should be made as simple as possible, but not any simpler." -- Albert Einstein
I think the downvotes are because your post is somewhat off topic.
OP responded to "Amazon.com is down" with "this is a lesson in over-engineering" - which it isn't, because Amazon.com is most certainly not overengineered for its purpose (I've seen the code with my own two eyes).
Your response is "not everyone needs extensively engineered systems", which is true, but is a non-sequitor from the previous posts.
FWIW, Obama and Romney have both lived abroad. I think it takes a certain kind of person to want to be president, and the effect of travel is different on a person like that.
For this and other reasons, I think electoral politics is one of the least effective places to put energy, in working for change.
You must not be flying the low-cost airlines in Europe. I have never seen any US-based airline that rivals the combination of low-price and bottom-level service provided by Ryanair. Unfortunately, the low-end airlines are so popular within Europe that it's increasingly hard to avoid them, even if you want to pay more, because they have many of the direct flights (especially to vacation destinations).
Unless your destination is as low-budget as the airlines, it's pretty easy to avoid them. They're mostly the equivalent of a cheap holiday bus.
Also, although I will avoid the EasyJets and Ryanairs as much as possible, for a direct flight the duration of a bus ride it doesn't really make that much difference.
I don't think this is true at all. Europe's low cost airline business is major. Easyjet, Ryanair, Flybe and Jet2 just to name a few. Obviously the service on these carriers is awful but flying between European cities for $50 a flight isn't uncommon. Personally my most consistent good service experiences have been with Air Canada. BA was good too as was Aer Lingus. I've only had two flights with them but I wasn't fond of AA.
| Personally my most consistent good service
| experiences have been with Air Canada.
I recently saw a headline that Air Canada is the third worst international airline for arriving on time. That said, I had no complaints about my only Air Canada trip (3 connecting flights).
As a Canadian I've used AC and other regional carriers quite a bit. Have to give a shout-out to Porter, if you're flying to the few cities in North America they service, they're top notch. Especially flying into Toronto, they exclusively use the Toronto Island Airport (YTZ), which is much more pleasant and better situated than Pearson (YYZ).
I think AC has gotten better, at least on domestic flights, since they've had more competition.
This was my only issue with Air Canada: my first and only (to date) flight with them was several hours late. It wasn't a problem for me so I think of them as being one of my preferred airlines because of how good everything else was. If the lateness had been a problem I might have seen it differently, though I didn't think the delay was their fault.
I've flown from Ireland to the US and Canada with them several times now and the service has been excellent. Now that you mention it though most of the flights arrived slightly late (no more than 30 mins). As they were trans-atlantic flights and I wasn't in a hurry it didn't bother me.
Even the US-based "legacy" carriers find themselves directly competing with Southwest, which has forced them to gravitate towards the low-cost model. The only thing that differentiates them in such an environment is their award systems. Except Southwest which (ironically?) has the best service and perks now.
It's admittedly been a few years since I flew a US low cost airline. But at least then they where pure luxury compared to Ryanair or Easyjet in Europe.
I compared US-based carriers with non-US based carriers. Southwest, a US-based carrier, competes with other US-based carriers. You're right, Southwest is a great choice if you're flying in the US, but that's not what my post is about. You can't compare Southwest with Cathay Pacific or EVA Airlines because they operate totally different routes.
Same thing regarding EasyJet comments, which is another "niche" airline. Try flying EasyJet from LAX to NRT, for example.
Alaska is pretty decent also, but I usually fly in the west coast market. Southwest prices are often not that great, actually, but their service is always good.
I've been badly screwed over by Leftouthansa many times, their response: "Nope, we don't care. We look forward to seeing you onboard." [After 2 support tickets, one to complain and then the other to report the unanswered first ticket after 2 months]
The market is capitalized at $2.5 billion, but what does that mean in an economy with no production?
Does the author define "production" as manufacturing where the exclusive currency used (for raw materials, wages, utility bills, etc.) is bitcoin? If so then I'm not sure of the relevance.
One could make all sorts of clever but offtopic points about a fiat paper currency when viewed through the lens of a P2P currency like bitcoin. "The Canadian dollar is supposedly worth more than the USD today, but what does that mean when
the Canadian network has zero nodes?" The wrongness of such a statement leaps out since fiat currencies are familiar to us. In contrast, bitcoin doesn't need a manufacturing base, but this is less obvious since P2P currency is a new idea
that people (myself included) don't well understand.
No, production is not exclusive to manufacturing. It means valued-added. For example, if you bought a dvd from amazon using bitcoin (I have no idea if this is possible), then very little of what you paid would count in bitcoin GDP. All of the costs of production (warehouses, delivery, labor, servers, etc.) are done in the US economy using dollars, so that basically doesn't register as bitcoin GDP.
On the other hand, let's suppose someone is buying a bag of... something. Well, everything that got that bag to the dealer was done in some other production economy. So the only value added, that would count as a service in bitcoin GDP, would be the dealer's profits, measured in bitcoins.
To my knowledge, there are no investment projects being financed in bitcoins. If there were, then they probably just defaulted on their loans, which increased tremendously under deflation. Therefore, it's best to think of the bitcoin economy almost entirely as pure exchange, with very little value added and very little wealth created.
As for fiat currencies, why is everyone so interested in the nominal exchange rate anyway? This was a peculiarity that I noticed a while ago. In college, I went on a study abroad program to Mexico. When we arrived, we all exchanged our dollars for pesos. Everyone was carrying on about how strong the dollar was, because one dollar could buy ten pesos. But then I noticed something. I went to McDonalds, and instead of a "dollar menu", they had a "10 peso menu". Who cares how many pesos you get for a dollar? It's about how much you can buy with that dollar worth of pesos, and that's measured by the real exchange rate. I realized at that moment that purchasing power was much closer to parity than most people think, because they put so much weight on nominal rates.
Recently I was offered a free flight for a weekend trip. I declined, citing the myriad displeasures of flying.
Now, should we initiate a behavior recall? Take the number of McDonalds-level staff in the air (A), multiply by the probable rate of power tripping (B), multiply by the average loss of tickets sold (C). A times B times C = X. If X is less than the cost of some basic decency training, we won't do any. Which airline do you work for? A major one.
In any sufficiently large organization, the people who make the decisions and the people who enact those decisions are separated by distance, background, experience, or other reasons. This is why you see so much dysfunction in large organizations, and it extends to politics and other things that affect you as well. Look for it and you'll start seeing it everywhere.
The Open DNS Resolver Project has a list of 25 million open resolvers. You can query their database for your IP address or up to a /24. Their site also has information on how to reduce or eliminate the problem via a couple options (RRL, BCP38). If you run a BIND resolver, consider switching to unbound. Part of this problem is rooted in BIND combining resolver and authoritative service in one daemon, which IMO "mis-educated" a lot of people.
Once you got black-listed getting removed wasn't always an easy process no matter how quickly you tried fix whatever caused it.
I took a job in the year 2000, at a company with 3000 email users, listed by Spamhaus. First thing I did was close the open relay they were running. The listing was promptly removed, and the mail queue was back to normal within only a few days. I'm skeptical of your claim. I've never seen a confirmed case of Spamhaus aggression, but I've seen a lot that were disproven, and even more that sound like they were written by miscreants. Like the kind who would advocate DDoS attacks cough.
Open relays were at the time manageable, even the ones that suddenly appeared when someone installed an old OS-version, as were the process for getting removed from the blacklists due to open relays.
Once you had one a computer lab workstation hacked and used for spamming - not so easy to get whiteliested anymore.
The university had a class B-network, trying to get the staffs subnet whitelisted while keeping the computer-labs blacklisted was apparently not possible according to the spam clearing houses. Blocking port 25 for outgoing traffic not possible to check from the outside and didn't help.
I can understand that organizations like spamhaus are overworked don't have the resources to handle every non-standard case on the internet as quickly as the blocked ip-range would like, but the replies we got were truly unhelpful.
The fact that someone bothered to register the domain stophaus.com seems to indicate that my experiences isn't uniqueue.
Furthermore, there's a lot of FUD in this thread about Spamhaus listing people who don't emit spam. IF this is true, then Spamhaus would have an unacceptably high false positive rate, and we would be able to observe this. In reality, Spamhaus has the lowest FP rate in the industry. Occam's Razor suggests those who claim to have been wrongly blocked are mistaken about the reason for their listings (if they ever existed in the first place).
You are incorrect. (Well, you're correct that Spamhaus doesn't filter content -- but they don't filter anything, they publish lists that various filtering software uses.)
I hear the SBL can also block domains, how? What is "URIBL_SBL"?
Yes, the SBL can also be used as a URI Blocklist and is particularly effective in this role. In tests, over 60% of spam was found to contain URIs (links to web sites) whose webserver IPs were listed on the SBL. SpamAssassin, for example, includes a feature called URIBL_SBL for this purpose. The technique involves resolving the URI's domain to and IP address and checking that against the SBL zone.
And of course they also have the DBL (Domain Block List), though I don't know if that existed back when PG ran into problems.
Do you have a link to the false positive rankings? I'm curious as to how that is measured.
Good point; I think both of our statements are true due to ambiguous wording upstream. I also took it literally, "any filter relying on the SBL" -- I use the SBL (via ZEN) but don't use SpamAssassin. And so my mail servers wouldn't block any domain that resolves to an IP address in the SBL, as described in the link you provided.
As for DNSBL false positive rates, I haven't seen statistics in a few years, and by now they wouldn't be worth much. The only ones I saw were from 2005 or 2007. This one (linked to from the below article) from 2011 doesn't even test Spamhaus:
This is just my personal experience saying (in 2013) that Spamhaus has the lowest FP rate, which isn't scientific. I'm kind of surprised there haven't been more FP comparison reports of major DNSBLs in recent years. If anyone has a link I'd love to see it.