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

Self-Hosting Still Pays

servethehome.com

Post reply on HN
Print view View original

301–310 of 343 posts

Re: Self-Hosting Still Pays

#301

Post by kroolik »

anonfornoreason wrote: post #222
kroolik wrote: post #108

Earlier quoted context omitted.

You and people posting similar answers miss a very important point. You've mentioned you have over 10 years of experience managing rack servers. That's super great! This means a company that hires you does not need to use AWS for some of the services, because you've got the necessary expertise to do it yourself. Other companies buy this expertise via fully managed services, like S3 or Aurora. Once you have expertise…

This is a startup/business/tech forum, I am a founder of a 20 person business (hiring 7 more) that does this part-time because it's very low maintenance. I am finally hiring a devops person to handle the command line side of things, though I am still pretty happy to dink around at the datacenter every once in a while. I taught myself as I went. I save about 10x over AWS. I would likely make the same decision today if…

I'm glad you've taught yourself infra things as you needed. But, you see, this is an anecdotal proof (like the whole article, anyway).

Learning infra is not the case all the time. Some companies choose AWS (or any other) for speed.

We can, again, argue but the argument is my PoV vs yours, or any other new commenter. Although I know how to maintain small infra: bunch of vms, reverse proxy and an rdbms, I'm not really into hosting HA Rabbitmq or Postgres myself. There is a ton of problems and configuration options coming from replication and consistency I would rather pay premium to people who have done it in the past.

I want my team to build solutions solving customers problems instead of fixing bugs in our replication code. It's all about perspective, I guess we can't reach a definitive answer for one or the other.

Re: Self-Hosting Still Pays

#302

Post by busterarm »

wolco2 wrote: post #276
busterarm wrote: post #242

Earlier quoted context omitted.

Here we are in agreement. Thing is though, people who well-understand AWS/Azure/Google are plentiful (but expensive) in the industry at this point. Small startups can't afford to have junior talent.

Small startups use junior talent for most services. They are not hiring the best hr team, marketing, sales and development. They don't have unless amount of money. They will pick one or two areas where they have an advantage and hire a strong person or two.

I think we have wildly different expectations of the scale of "small startups".

If you can afford to full-time hire an entire hr team, an entire marketing team, an entire sales team and an entire engineering org, then you're able to pay an AWS bill.

Re: Self-Hosting Still Pays

#303

Post by busterarm »

rualca wrote: post #295
User23 wrote: post #206

Earlier quoted context omitted.

Amazon itself used to be that environment. Before AWS was adopted internally, which it wasn't for some time, to say nothing of the many years before it was created, getting capacity involved a weeks or even months long political game with the infrastructure team. Even infrastructure devs had to struggle to find resources. Hiring managers would literally horse trade their spare capacity.

> Amazon itself used to be that environment. Before AWS was adopted internally, which it wasn't for some time, to say nothing of the many years before it was created, getting capacity involved a weeks or even months long political game with the infrastructure team. It's my understanding that you're describing how Amazon worked in the 90s. It's my understanding that Amazon afterwards moved to an internal cloud provide…

> It's my understanding that you're describing how Amazon worked in the 90s. It's my understanding that Amazon afterwards moved to an internal cloud provider, and afterwards spun off AWS which worked in parallel with Amazon's internal cloud provider.

Not at all. Amazon's transformation took years. Up until last year, amazon.com was Oracle's biggest customer. AWS development far outpaced their actual use of it.

Re: Self-Hosting Still Pays

#304

Post by fomine3 »

ocdtrekkie wrote: post #252
sokoloff wrote: post #189

Earlier quoted context omitted.

O365 had a couple of bad years, earning the nickname "Office 360" for a while. This year, I can only recall a single outage and we've been overall happy with it. O365 costs more than an Exchange server only in part because it comes bundled with all of MS-Office.

If you bundle Exchange + CALs + Office Professional Plus, over a three-year version upgrade cycle, versus three years of Office 365, Office 365 costs about twice as much, last I checked. And, since Microsoft supports it's software for ten years, you can skip one or two versions of Exchange/Office and still get security updates, and pay a quarter or a sixth of what it costs to go with Office 365.

For comparing email service, cloud services are superior for spam reputation because they are big.

Re: Self-Hosting Still Pays

#305

Post by fomine3 »

jacobsenscott wrote: post #215
drchopchop wrote: post #47

Earlier quoted context omitted.

Also, when it comes time to scale, it's a lot easier to bring up a relatively simple (but scalable) infrastructure on AWS than deal with: colo location, buying servers, buying spares, replacing failing hardware, imaging, setting up backups, etc etc. Those typically become concerns once you're at big scale where the cost savings are significant.

But for most companies it never "comes time to scale". Especially if you have 100% dedicated hardware and you can start out with a 64 core db server with 120GB of memory and a matching web server pretty much for free. You need to "scale" early in a cloud environment because you are allocated a fraction of the hardware and someone else on that same hardware is using all the resources. AWS/DO etc boost their margins by…

VPS packs but EC2?

Re: Self-Hosting Still Pays

#306

Post by welterde »

ericpauley wrote: post #224
welterde wrote: post #207

Earlier quoted context omitted.

There is no need to have anyone on call yourself to drive to the colo, since any decent colocation provider will have staff on-site 24/7 to perform any physical intervention required. Part of that you pay through your colocation bill and part you pay by the hour (depending on your contract/provider).

They're doing their own routing. Will the colo tech fix their iptables if they mess them up and can't ssh?

Ideally one always has OOB (out of band) access to all devices to make sure one can always recover from any error, such as messing up firewall config or similar. But there are transactional network management tools that can automatically roll back changes in case one messes up.

But the colo techs can do pretty much anything you would do yourself.. Plenty of people run racks in locations they have never set foot into themselves..

Re: Self-Hosting Still Pays

#307

Post by busterarm »

dragonsh wrote: post #275
busterarm wrote: post #248

Earlier quoted context omitted.

> If you are working with AWS your application and your whole organization is at the whims of Amazon. Part of freedom is the freedom to do stupid things. > You will not even be able to migrate if you are tied into those API's and services because they are not same across cloud providers. VPCs, Instances, Load Balancers and Storage are roughly-equivalent everywhere and the differences between are mostly-trivial. For t…

> VPCs, Instances, Load Balancers and Storage are roughly-equivalent Worked enough as preferred partner with all 3 major cloud providers to know that roughly equivalent does not mean equal and efforts are significant. e.g. self-hosted haproxy or nginx is much better than those load balancers as each provider has their own set of API's. So please do not underestimate the efforts. > If you think this doesn't also apply…

You can own your data and still use public cloud.

You are familiar with using multi-site DR strategies, yes?

Re: Self-Hosting Still Pays

#308

Post by didibus »

joana035 wrote: post #211
didibus wrote: post #148

Earlier quoted context omitted.

When you say AWS? You mean EC2? Cause when I think AWS, I think of using Lambda, Fargate, DynamoDB, RDS, SQS, SNS, S3, Kinesis, Redshift, Elastic search, etc. All these managed services arn't just about cost or scale, a lot of it is about convenience and development speed and flexibility, as well as reduced maintainance and operational burden.

"maintenance and operational burden" in the cloud is way higher than renting a box somewhere and forgetting about it. I can't understand why people keep using this as an argument against anything else other than cloud while playing catch-up with the rain of changes and suffering from fomo of last week's new stuff from cloud. Heck, when did infrastructure became like js landscape?

Doing OS updates, configuring load balancers, DB backups, setting up firewall rules and OS user accounts, ssh access, archiving logs, configuring monitoring and alarms, assigning DNS, handling deployments, etc.

I don't know, as an application developer, I'd rather not waste my time with stuff like that if I can avoid it.

Re: Self-Hosting Still Pays

#309

Post by dragonsh »

busterarm wrote: post #307
dragonsh wrote: post #275

Earlier quoted context omitted.

> VPCs, Instances, Load Balancers and Storage are roughly-equivalent Worked enough as preferred partner with all 3 major cloud providers to know that roughly equivalent does not mean equal and efforts are significant. e.g. self-hosted haproxy or nginx is much better than those load balancers as each provider has their own set of API's. So please do not underestimate the efforts. > If you think this doesn't also apply…

You can own your data and still use public cloud. You are familiar with using multi-site DR strategies, yes?

> You can own your data and still use public cloud.

First it will be good if you read through the terms of service, privacy, data retention and service agreement contract. You will be astonished to know that buried deep within the legal language the data is owned by the cloud provider (as they own access as well as the rights to physical resource holding the data) and they provide you an option to export it (but this can be completely blocked by a notice from govt.). You can mitigate the risk by encrypting the data with your own key which is not shared with cloud provider (but indeed this is also not true most of the time given the key to encrypt the data is hosted with the cloud provider as well).

For your reference AWS for its instances still do not support ed25519 key for SSH access and needs an RSA public key (although its less secure). We need to go extra mile to customize the AMI images to support ed25519 keys and remove the account and keys created by AWS.

Please let me know how you will recover the data in case your multi-site deployment on AWS with DR is blocked by aws as they received a notice from government agency with a gag order.

> You are familiar with using multi-site DR strategies, yes?

Most companies using AWS or GCP or Azure use multi-region deployment (which means using different regions). But all will be shut down with a one single notice from government agency with gag order (you won't even know why given gag order).

So your multi-site DR will only work if the company keeps a self-hosted backup of data to avoid cloud provider from shutting down access to the data. This is precisely my point that self-hosted guarantees freedom not the cloud which is right opposite of freedom (traded for convenience).

Re: Self-Hosting Still Pays

#310

Post by oblio »

dragontamer wrote: post #168
oblio wrote: post #126

Earlier quoted context omitted.

Tftp?

Tftp is commonly used for PXEBoot scenarios. You have a centralized PXEBoot server (basically a DNS + TFTP Box). All servers you manage auto-boot to PXEBoot, then download their image from the central server, then automatically provision themselves through Puppet or Kubernetes or whatever the heck you're doing. Throw down a few network power-adapters (to turn on/off in times of emergency), IPMI for network remote KVM…

The centralized PXEBoot server is a single point of failure, at least how you presented it.

You also need expertise in DNS/(T)FTP/PXEBoot/hardware/Kubernetes/Puppet/IPMU/KVM/VPN/etc.

It's like that joke about the expert called in to repair expensive factory equipment. He comes in, takes a piece of chalk out of his pocket, marks the piece that should be replaced with an X. A month later they get the bill: $100000. The factory manager wants to see a detailed invoice for that huge expense. They get it 1 week later:

$1 for the piece of chalk

$99999 for the expertise needed to know where to put the X

Post reply on HN