Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Farewell EC2-Classic, it’s been swell

allthingsdistributed.com

Post reply on HN
Print view View original

131–140 of 187 posts

Re: Farewell EC2-Classic, it’s been swell

#131

Post by icedchai »

saurik wrote: post #115

EC2-Classic was magical. It felt like we were finally living in the future: a world where software was in charge of networking, and all of the legacy cruft that came from having to build networks out of cable could be forgotten. Rather than care about legacy IP "subnets", the cloud cared about "security groups", which were missing only a couple features (such as "hierarchy") to entirely replace the role of a subnet i…

One of the first startups I worked for was all on EC2-Classic. I did enjoy its simplicity. I can understand the need for VPC when integrating with on-premise networks, VPNs, etc. However, you often run into cases where VPC is simply not necessary and over complicates things.

Re: Farewell EC2-Classic, it’s been swell

#132

Post by recursive »

xnx wrote: post #67
Kuinox wrote: post #17

Earlier quoted context omitted.

I started to use Kagi instead of google search. Since, it's difficult to get back on google, the results are so much worse.

What's an example search that's better on Kagi?

"Radiator". I just picked a word at random. All the other ones are better too.

Re: Farewell EC2-Classic, it’s been swell

#133

Post by wombat-man »

lolinder wrote: post #82
wombat-man wrote: post #69

Earlier quoted context omitted.

I think it just was never going to get the mass adoption it needed to justify the upkeep.

The upkeep was apparently 12 engineers and they had tens of millions of users. It doesn't take that many users to justify 12 engineers plus infra, so it sounds like it was more that Google doesn't care to operate any product unless it will have users in the hundreds of millions.

I don't think they ever put ads on it, really wonder what the decision process on that was.

Re: Farewell EC2-Classic, it’s been swell

#134

Post by nine_k »

kibwen wrote: post #127
sneak wrote: post #9

Earlier quoted context omitted.

I miss Google Search being good and useful and not covered in intrusive ads much more than I miss something easily replaced with local software or Feedly. Almost everything Google does outside of GCP, Maps, Search, and YouTube could evaporate for all I care. Google's problem is not that they cancel stuff, it's the everpresent need to grow revenue and embed annoying ads into more and more of everyone's daily lives. I'…

> everything Google does outside of GCP, Maps, Search, and YouTube could evaporate for all I care Honestly, even YouTube could evaporate and it really wouldn't make a difference. 99.9999% of YouTube is just mindless entertainment, which is 100% fungible with every other form of entertainment. The amount of actually unique, insightful, worthwhile content on YouTube is a rounding error, and will find other places to li…

AFAIK more than 95% of all email traffic is spam. I still consider email an absolutely vital tool, despite the efficiency below a steam engine.

Same with YouTube: the small sliver of content I care about is important enough for me to pay for YouTube premium.

Sometimes people compare something to a gold mine, to emphasize how rich that is. A typical gold mine extracts several grams of good per tonne of rock, that is, a few parts per million.

Don't cry about the Sturgeon's law; embrace it and celebrate what you can extract.

Re: Farewell EC2-Classic, it’s been swell

#135

Post by jedberg »

Bluecobra wrote: post #62
brianshaler wrote: post #55

Off topic, but I'm confused about the IPv4 notation in this quote: > When we launched EC2 in 2006, it was one giant network of 10.2.0.0/8. In my understanding, /8 means the first octet would be fixed (10.0.0.0-10.255.255.255) and I'm having trouble understanding the significance of the 2 here. If the 2 is significant, wouldn't you write it as /16? Given the source and my confidence on this notation, I feel like there…

I think it must be a typo. because 10.2.0.0/8 is not a valid subnet mask. The largest subnet you can make starting with 10.2.0.0 is 10.2.0.0/15.

It's still a valid subnet notation, the 2 just doesn't mean anything.

When you take the logical AND of the IP and the expanded mask, you'll just end up with 10.0.0.0 as the network address and 10.255.255.255 as the broadcast address.

Re: Farewell EC2-Classic, it’s been swell

#136

Post by jedberg »

saurik wrote: post #115

EC2-Classic was magical. It felt like we were finally living in the future: a world where software was in charge of networking, and all of the legacy cruft that came from having to build networks out of cable could be forgotten. Rather than care about legacy IP "subnets", the cloud cared about "security groups", which were missing only a couple features (such as "hierarchy") to entirely replace the role of a subnet i…

If you create a new account, it will work like classic EC2. They will set up the VPC for you behind the scenes. Until you "break the glass" and try to configure a VPC, it will work just like old classic did.

Re: Farewell EC2-Classic, it’s been swell

#137

Post by werner »

brianshaler wrote: post #55

Off topic, but I'm confused about the IPv4 notation in this quote: > When we launched EC2 in 2006, it was one giant network of 10.2.0.0/8. In my understanding, /8 means the first octet would be fixed (10.0.0.0-10.255.255.255) and I'm having trouble understanding the significance of the 2 here. If the 2 is significant, wouldn't you write it as /16? Given the source and my confidence on this notation, I feel like there…

You are right, it was a typo. It should have been 10.0.0.0/8. Updated that in the blog, also resolved a confusion about the memory in a P3dn.24xlarge.

Thanks for your critical reading!

Re: Farewell EC2-Classic, it’s been swell

#138

Post by mike_hearn »

lolinder wrote: post #82
wombat-man wrote: post #69

Earlier quoted context omitted.

I think it just was never going to get the mass adoption it needed to justify the upkeep.

The upkeep was apparently 12 engineers and they had tens of millions of users. It doesn't take that many users to justify 12 engineers plus infra, so it sounds like it was more that Google doesn't care to operate any product unless it will have users in the hundreds of millions.

By the end it had an upkeep of one dude's 20% time. It was very reliable and the Google infrastructure didn't require much ongoing maintenance at that time, the servers just kept on trucking.

Re: Farewell EC2-Classic, it’s been swell

#139

Post by mike_hearn »

wombat-man wrote: post #133
lolinder wrote: post #82

Earlier quoted context omitted.

The upkeep was apparently 12 engineers and they had tens of millions of users. It doesn't take that many users to justify 12 engineers plus infra, so it sounds like it was more that Google doesn't care to operate any product unless it will have users in the hundreds of millions.

I don't think they ever put ads on it, really wonder what the decision process on that was.

Many blogs have ads on them. So you'd be stripping ads and replacing them with your own - not cool.

Re: Farewell EC2-Classic, it’s been swell

#140

Post by MichaelZuo »

snewman wrote: post #109
MichaelZuo wrote: post #39

Earlier quoted context omitted.

This. Unless a written guarantee could be offered by the cloud provider that migration will be absolutely trouble free, why wouldn't customers just stick to what they know works? And if they're not willing to provide a written guarantee, then that says a lot.

In this scenario, it's nothing to do with the cloud provider. Migrating a live production system is inherently difficult. You can make a reasonable analogy to moving houses – say, with two kids who are in school, and while you and your spouse are both working. No matter what guarantees you're given regarding the condition of the new house, simply moving all of your stuff (while you are using it) is a big hassle.

In this example, it wouldn't be the guarantees for the new house, which presumably would have been examined and accepted well beforehand, it would be guarantees for the moving process itself.
Post reply on HN