Hacker News Reader

Comment by zelon88 | Hacker News Reader

> PaaS is always going to offer less for the same price, but what surprised me was how start the difference can be. On Render.com, a 4GB RAM and 2vCPU service costs $85/mo. The same spec costs $14/mo on Scaleway (ARM).

I notice a trend that the people who scoff at hardware specs are usually the same ones standing in line for 2 cores and 4gb of RAM for $50+/month. They'll laugh when you suggest utilizing an obsolete $50 computer with a decade old CPU (that's just sitting in a closet), but more than willing to spend $50/month on similar performing hardware from a Cloud vendor.

View on HN

Cue the downvotes but I imagine you could sort those people into two buckets:

1. Those who love their Macs a little too much

2. Those who routinely build a PC.

i.e. Are you familiar with the hardware market.

I'm not gonna downvote you. But to your point about the hardware market, a mid range server CPU of today will be about as fast as an entry model CPU next year. Actual performance wise, your old server can probably keep up with a newer but slightly lower end server just fine. Obsolete !== useless.

I'm the kind of person who would rather take the 3y server and recomission it as a lower priority service than just do a 1:1 replacement. "Old" computers aren't as useless as they used to be. Computing power has advanced to the point where computation is arbitrary. For all intents, in most sectors, you can scale your compute capacity as high as your budget allows and you won't hit any performance barriers ever. There will always be more compute. It is a commodity now.

My stance is this; sure the new server is faster than the old one, but you know what's even faster? Dividing the existing workload between both of them.

> sure the new server is faster than the old one, but you know what's even faster? Dividing the existing workload between both of them.

How do you factor reliability into your analysis? I mean, the reason companies ditch their 3yo server is that their service life is nearing it's end, and odds are they will fail in the upcoming year or so.

Meanwhile, increased compute capacity means companies can use a single brand new server to replace multiple old ones, not to mention cheaper operational costs.

> a mid range server CPU of today will be about as fast as an entry model CPU next year

What about the core count and memory capacity

https://www.amazon.com/Dell-PowerEdge-R630-Rackmount-Warrant...

I think people know they’re paying a premium on hardware. The point of PaaS isn’t to get the best deal on specs it’s for small teams to iterate quickly and focus on product. If they’re successful they’ll outgrow it and hire people to manage infra.

3. Those who build their own hackintosh because they love OSX too much.

It is not the hardware actually. It is network reliability makes cloud better.

I've seen real-world connectivity, latency, and bandwidth problems crop up with enough frequency to be a real problem on a major budget "cloud" provider. It looked liked they'd badly cheaped out on their peering agreements.

Move the exact same workload to the industry's default but much more expensive choice, and the problems vanish entirely.

This is, unfortunately, one of those things that's really hard to judge about a hosting provider unless you have direct experience using them "at scale", as they say. Nobody puts that stuff on a sales page spec sheet or comparison grid.

Could I save money by hosting on real hardware at some popular, cheapish server-leasing place? Or colocating at one? Or hosting out of my own basement!? Maybe. Would it cause some users to consistently see dial-up speeds and dropped connections on gigabit Internet service because of either some quirk of routing, or bad peering agreements? Who knows!

I’ve observed those issues on the industry default CSPs. Terrible latencies for some services. Some of those budget CSPs simply resell AWS and Azure and GCP and OCP, while others run on commodity VPS infrastructure. And honestly there’s still too much ops work running workloads not on-premise, wasted time even though totally entertaining sometimes.

Not just that, it's also all the legwork of managed hosting: all that OS configuration, patching, redundancy, testing, monitoring etc etc is someone else's job, and they are accountable. Plus other managed services like LB, DB, auth, etc etc. you might not want to duct tape yourself and manage every day. That cloud bill could be cheaper than your time is worth.

Plus flexibility of scale up / scale down as needed for load, transient testing, etc.

Cloud is definitely not for everyone but it makes sense for some.

> that OS configuration, patching, redundancy, testing, monitoring etc.

For larger projects I guess, but I have a few VPSes that have been running sites for a decade or more that require almost no maintenance (occasional apt-get upgrade). In fact it’s the frameworks / languages these projects were built in that cause most of the work (for which deployment target makes no difference).

With IaC it’s become feasible. You kinda iteratively build it up. With something more predictable like Nix it really gets close to shops run by a single person. And on e you start to get managed infrastructure, next question that pops up would often be which services do you want to delegate, and some folks wind up in serverless land—locked in after a while, which might be totally fine.

Colocation exists. It's a solved problem. It's just not in-vogue right now.

and for $400/mo you can put your $50 old machine in someone else's closest.

I am renting 16 core AMD with 128GB RAM and 2x2TB SSD for less then $100/month in Canadian pesos on Hetzner.

I also pay for 1Gbps symmetric fiber with dedicated IP going into my house ($140/month at the moment) so I also host some stuff right from my own place.

And with my $50 old machine I can host $5,000/month worth of "cloud" services.

Hetzner & OVH gets you the best of both worlds.

Unbelievable (new) specs, hosted and at prices points tough to beat doing it yourself.

E.g.

Hetzner

  AMD RYZEN™ 9 7950X3D (current gen Zen 4)

  16 physical cores (4.2 GHz)

  128GB DDR5 ECC ram

  4TB NVMe

  completely hosted

For only ~$100/mo

https://www.hetzner.com/news/new-amd-ryzen-7950-server

Both are quite happy to terminate users’ accounts, but on paper it’s really compelling indeed.

Have you seem them terminate unreasonable situations?

They normally are trying to keep their IPs and Network clean of scammers abusing their resources, which inevitably hurt all customers.

My 2024 goal was creating a website for me and hosting a copy of my newsletter there. I heard a lot of good things about Hetzner and choose them. Just right after signing up, even after I put a credit card with 3DS they immediately banned me. I wrote to support, they forwarded me to page. I uploaded my passport with my face. But my account is still banned. And worst thing is I’m 1 month late for my 2024 goal.

Try racknerd… they have various Black Friday/new year specials that rarely/never expire and the price can be as low as $25/year for a decent 2-cpu VPS. Not the same hardware as OVH, storage often limited, but still an interesting price point and service. Biggest limitations: no 2FA on the control panel and mounting an ISO requires talking to customer service… oh and no automated backups but they might have recently added snapshots.

I should add - keep your own backups. I can’t trust that a service this cheap will not lose my data somehow. It hasn’t yet, but with margins this thin I wonder if they’ll ever go out of business on me.

> I wrote to support, they forwarded me to page. I uploaded my passport with my face. But my account is still banned.

Is there still more needed to complete their KYC process?

Hetzner can and will terminate your account and delete all your data on a whim, never telling you why. There is nothing at all you can do to prove your usage was legitimate. I get that they must deal with a lot of abuse but I wouldn't use them for anything mission critical ever.

Any reliable provider? Or is the solution to buy a server and colocate somewhere else?

You get what you pay for. You own what you possess.

AWS is expensive and not without trouble but in my experience they’ve been best of the pack.

Otherwise, on-premises + collocation is never a mistake if the maintenance and upright expenditure is in your budget.

> We’ve mounted the AX102 with 128 GB of DDR5 ECC RAM. In addition to the classic ECC (error correction code), which protects data both on the memory module and during transfer, the new DDR5 memory generation uses the on-die ECC method. This method carries out independent error correction directly on the DRAM chip, giving you greatly optimized reliability and data integrity.

Isn't on-die ECC necessary just because DDR5 is less reliable than previous generations, especially at higher frequencies and density? Makes me wonder whether they use an actual ECC DDR5 memory.

Have you used them? Horror stories on their billing / admin side (ie unilaterally closing your account and being unreachable) abound…

Yet many other folks praise them…

What’s your experience?

With comparable reliability, resilience, durability, and latency? I doubt scalability up or out is reasonable to ask, correct me if I’m underestimating its potential.

If we need to scale up, we drive to the nearby computer store, give them several thousand dollars for a bigger box, go home and install it!

I colo an internet search engine for $100/mo. That's power + network + rackspace.

I have 14core/128GB RAM, 2x900GB nvme + 60TB spinning rust with 33TB traffic in someone else's closet and paying only $60 ¯\_(ツ)_/¯. Prices vary a lot when it comes to colocation

That 60 for collocation only, right? How are you amortizing the machine if that’s the case? What SLAs with the closet owner?

yeap - 60 for collocation only. 99.9% uptime SLA (power/network). I've had 5-minute blips 3 times in last 12 months. The hardware itself is bought used $400 for the whole server (not including HDD). HDD was around $400 too. so last 2 years: (24 * 60 + 800)/24 = ~$94/month and going down every month

Yup, I did this (not for $60!!) almost a decade ago with 4 savemyserver HP 1U’s and a switch. Each server cost me about $200 without HDD. Another $200 in ECC memory. I used it for a long time until colocation costs went up because they were bought by L3Harris.

I run a couple ryzen mini PCs on Verizon fios 1Gbps at home now for my hosting needs but I would jump at $60 colo in a heartbeat if I got even 100mb connection unmetered.

That’s totally efficient, awesome, thanks for sharing.

It would be great if you could tell us the name of the colocation provider. I would really like to explore it.

This is a bit exaggerated. I've had a 1U or 2U server box colo'ed for the last 20+ years, and it's been consistently in the $80-$100 range. Double my total costs if you add that I'm replacing a $5K server box every 5-7 years or so.

The duct-tape-maintenance vs your-time criticisms are 100% dead-on though.

Depends on the industry I think

But isn't it more than just the vendor's hardware that one is paying for. Namely, the vendor's internet connection. Apologies if I am missing something obvious.

And for network redundancy / UPS / datacenter security / etc. Comparing a PaaS to some old PC running in a closet is completely missing the point.

If the thought of your service going down for a few hours because some random unplugged the power cable or your ADSL router crashed doesn’t make the CFO lose sleep then some raspberry pi is surely good enough. Just make sure you run it inside a safe if you store personal information.

Reply on news.ycombinator.com