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

Why we use our own hardware

fastmail.com

Post reply on HN
Print view View original

421–430 of 547 posts

Re: Why we use our own hardware

#421

Post by brongondwana »

rob_c wrote: post #63

Hosts online service seems to think deserving of medal for discovering that S3 buckets from a cloud provider are crap and cost a fortune. The heading in this space makes your think they're running custom FPGAs such as with Gmail, not just running on metal... As for drive failures, welcome to storage at scale. Build your solution so it's a weekly task to replace 10disks at a time not critical at 2am when a single disk…

Ceph is most certainly not the only game in town. It's good and stuff, but it's just tech. We're using protocol level replication for each of our data stores.

Re: Why we use our own hardware

#423

Post by fireflash38 »

Xylakant wrote: post #346
evantbyrne wrote: post #252

Earlier quoted context omitted.

KMS access is granted by either environment variables or by authorizing the instance itself. Either way, if the instance is compromised, then so is access to KMS. So unless your threat model involves preventing the government from looking at your data through some theoretical sophisticated physical attack, then your primary concerns are likely the same as running a box in another physically secure location. So the sa…

An attacker gaining temporary capability to encrypt/decrypt data through a compromised instance is painful. An attacker gaining a copy of a private key is still an entirely different world of pain.

It's the same pain, since the resolution is the exact same. You have to rotate.

Re: Why we use our own hardware

#424

Post by watsocd »

ttul wrote: post #329
johnklos wrote: post #65

The whole push to the cloud has always fascinated me. I get it - most people aren't interested in babysitting their own hardware. On the other hand, a business of just about any size that has any reasonable amount of hosting is better off with their own systems when it comes purely to cost. All the pro-cloud talking points are just that - talking points that don't persuade anyone with any real technical understanding…

My firm belief after building a service at scale (tens of millions of end users, > 100K tps) is that AWS is unbeatable. We don’t even think about building our own infrastructure. There’s no way we could ever make it reliable enough, secure enough, and future-proof enough to ever pay back the cost difference. Something people neglect to mention when they tout their home grown cloud is that AWS spends significant cycle…

...and power, backup power, HVAC, physical security...

Re: Why we use our own hardware

#425

Post by beaugunderson »

I don't trust anything from fastmail after they bought pobox and forced me onto their new service which fails at the one thing pobox did well--forwarding email. They also refused to give me a refund (prorated or not) for removing the product I was using and substituting a defective one.

Re: Why we use our own hardware

#426

Post by amelius »

jorams wrote: post #362
amelius wrote: post #305

Earlier quoted context omitted.

I guess what's confusing is that I'm calling the promotional emails also "spam". But thanks for your suggestions. I see a few problems. When I receive a promotional email, I want to add a rule, and I have to click 7 times (including once for "Archive"), and use the scroll-wheel to select the "Promotions" label. Secondly, the rule is not applied directly. This is confusing, and cumbersome. Note: I don't want to Unsubs…

One rule that may be good, depending on how much of the mailing list email you receive is promotional, is to match "A header called List-Unsubscribe exists" and move that to Promotional. Then you could put any exceptions that it categorizes wrong above it.

That's a good idea. Although, oddly, some of the emails that have an Unsubscribe button have no List-Unsubscribe header.

How would you suggest to solve the following problem: let's say I have archived all my mail (inbox zero); how do I now see the emails that are important to me (i.e., everything that was not labeled e.g. with promotions)?

Re: Why we use our own hardware

#428

Post by resonious »

rakoo wrote: post #391
swiftcoder wrote: post #198

Earlier quoted context omitted.

> All the pro-cloud talking points... don't persuade anyone with any real technical understanding This is a very engineer-centric take. The cloud has some big advantages that are entirely non-technical: - You don't need to pay for hardware upfront. This is critical for many early-stage startups, who have no real ability to predict CapEx until they find product/market fit. - You have someone else to point the SOC2/HIP…

> You have someone else to point the SOC2/HIPAA/etc auditors at. I would assume you still need to point auditors to your software in any case

You do, which makes it very nice to not have to answer questions about the physical security of your servers.

Re: Why we use our own hardware

#429

Post by jph00 »

riezebos wrote: post #365
jph00 wrote: post #344

The original answer to "why does FastMail use their own hardware" is that when I started the company in 1999 there weren't many options. I actually originally used a single bare metal server at Rackspace, which at that time was a small scrappy startup. IIRC it cost $70/month. There weren't really practical VPS or SaaS alternatives back then for what I needed. Rob (the author of the linked article) joined a few months…

As a customer of Fastmail and a fan of your work at FastAI and FastHTML I feel a bit stupid now for not knowing you started Fastmail. Now I'm wondering how much you'd look like tiangolo if you wore a moustache.

Now I wonder what he'd look like without the moustache :)

Re: Why we use our own hardware

#430

Post by calmbonsai »

bluGill wrote: post #404
llm_trw wrote: post #373

Earlier quoted context omitted.

>As the fortunes of AWS et al rose and rose and rose, I kept looking at their pricing at features and kept wondering what I was missing. They seemed orders of magnitude more expensive for something that was more complex to manage and would have locked us into a specific vendor's tooling. But everyone seemed to be flocking to them. In 2006 when the first aws instances showed up it would take you two years of on demand…

Aws is useful if you have uneven loads. why pay for the number of servers you need for christmas the rest of the year? But if your load is more even it doesn't make as much sense.

This.

Plus bidding on spot-instances used to be far less gamed so if you had infrequent batch jobs (just an extreme version of low-duty-cycle loading), there was nothing cheaper and easier.

I've been out of that "game" for a bit, but Google Compute used to have the cheapest bulk-compute instance pricing if all you needed was a big burst of CPU.

It's all changed if you're running ML workloads though.

Post reply on HN