Hacker News Reader

Comment by peterbonney | Hacker News Reader

Am I alone in not seeing a paradox here? It's no different from a variety of things that make sense at some scales and not at others.

Seems similar to office space - a 5 person company will find the economics of coworking space compelling, a 500 person company will do better with a long-term lease with a commercial landlord, and a 50,000 person company might have their own property management team in-house. That doesn't create a paradox in the commercial real estate market, just different solutions for different needs.

View on HN

The paradox is that if the coworking space were the cloud, it would just grow with you, and easily house a 50k-strong company, while you'd be paying the same (high) rate per employee.

Why would this happen? Because what the company builds is a large and growing money-making contraption right inside the coworking space, also using parts of the building for critical functions, and the door is intentionally kept small. The only way to move that contraption is to dismantle it and reassamble in a much cheaper, purpose-built hangar, rebuilding some of the critical parts along the way. During all that time, the contraption would stop making money.

That's the beauty of AWS business model: it's a no-brainer for a startup to use it, but the startup grows, more financially efficient infrastructure options become unattainable because of the very high cost of the move. This is the best-executed vendor lock-in I know.

This seems like a well known trap that people would know how to avoid with a little planning, but I guess they have no one thinking that far ahead

There's also the argument that some of the cloud advantages (such as AWS proprietary services like SQS) simply aren't directly available as selfhosted replacements. You end up needing a whole new team to do HA/scaling for a core service.

There are off the shelf things for, say, S3 and ALB which are entirely workable, but once you start getting into more complicated stuff (like S3's new consistency semantics, or SQS) then you're looking at a whole small company (at a minimum) worth of additional work. It's a non-trivial expansion, even for a large org with lots of money.

You can avoid using these sorts of services to maintain flexibility/independence, but you lose out on their unique benefits. This isn't an accident. It's not like you're going to be able to selfhost an SES clone and get the same kind of deliverability percentages as AWS netblocks, no matter how many engineers you throw at the problem.

There are ways to engineer around this, but it’s going to depend on balancing risk management with business agility. Maybe you use Backblaze for object storage instead of S3. Maybe you use Cloudflare for CDN and Serverless at the edge. Maybe you use Mailgun or another SES competitor. Containers let you use managed k8s off the shelf anywhere. Lots of options for technologists, PMs, or CTOs to pick from.

Graviton2? Aurora? S3 global read-after-write consistency? Glacier? SQS's ability to scale?

Some of the goods you just can't get anywhere else, and they know it.

It can be done. Putting the right abstractions in front of these services makes it possible to migrate and replacements exist. Okay, granted, there will be complexity in self hosted infra but it is possible.

I don't really think anyone with anything less than an 8 figure budget can get the perf/watt of Graviton2. They're literally not available for sale - AWS had them fabbed and then kept all of them to rent out.

It's called Ampere Altra. Maybe you should have picked a different example like Aurora...

Have you tried? Us startuppy mortals can't buy those today, unfortunately. Maybe in a few years, when they trickle down from the people with 8 figure budgets.

Aurora is the one service that got us to migrate to AWS! Honestly it saved us 5 digits per month over another MySQLaaS. For us it was a very good deal. We could’ve dedicated engineering time to reduce db cost on the old provider but Aurora was drop-in awesomeness.

Hardware is difficult, for sure. The managed services / proprietary services is another story.

Having said that, and please keep in mind that I do not have an idea much about hardware so this is a guesswork; more of a question than a statement..., I do think that the Graviton price vs performance might not have a direct benefit outside of AWS. It's cool for AWS because they have less stuff to manage at that scale and it might be cheaper for them to buy and iterate because only they use it. It looks great to the clients because the perception is that the value is better. But AWS runs 80 AZs in 25 regions so it benefits them at that scale.

Realistically, even if one had couple of hundreds of racks, does it matter at that scale?

Anyone with less than an 8 figure budget shouldn't even begin to consider self-hosting.

Graviton2 - there are arm vendors that sell arm hardware

Aurora - tidb or vitess or cockroachdb or citus etc etc

Glacier - there is backblaze, also onprem vendors

Sqs - just use Nats

> There's also the argument that some of the cloud advantages (such as AWS proprietary services like SQS) simply aren't directly available as selfhosted replacements. You end up needing a whole new team to do HA/scaling for a core service.

It's not limited to some. You absolutely need a whole new team (or teams) to handle your infrastructure, your high availability needs, and also your security.

The nifty serverless offerings and other features are just nice-to-haves in comparison to the core infrastructure work that cloud providers put into your system to keep it running.

Just because cloud providers like GCP and AWS and Azure and etc have everything put together to let you setup your whole infra by running a small script that does not mean nothing needs to be done in order for that to work.

> such as AWS proprietary services like SQS

SQS is a message queue service. It's a bit weird to claim that as "simply aren't directly available as selfhosted replacements", a message queue is pretty basic stuff.

SQS is "just a queue service" the way that S3 is "just an object store" and the iPhone is "just a smartphone".

I'm super into self-hosted shared-nothing queue services, but what AWS is doing with SQS is anything but basic stuff.

> I guess they have no one thinking that far ahead

Small businesses don't care because they are too busy surviving.

Medium businesses don't care because why break what works?

Large businesses don't care because their managerial careers promote short-term priorities and marquee "transformation" projects which are buzzword-driven, and cloud is the current one.

So what is the problem exactly?

Paying more for less because of marketing.

Life goes on. It happens all the time. The trick is to learn and be smarter next time.

The ones thinking that far ahead aren’t using AWS in the first place.

Yep, there the ones with stressed out Ops staff running around replacing hard drives at 3 AM and with managers trying to work out how they are going to pass the next security assessment because they don't handle end-to-end encryption or have sufficient controls in place.

Is this common?

My company recently shifted from on prem to cloud. And we aren’t primarily a software company (although, much like most companies, we are increasingly turning into one).

We outsourced our data center needs, and I don’t remember us ever facing these middle of the night massive breakdowns (we had load balancing between a couple of data centers) and the number of ops issues hasn’t reduced a bit since the switchover.

I do think it’s been easier to improve performance for non US customers, because spinning up servers in data centers across the world seems a little easier around the world, but the millions saved by self hosting would have easily paid for someone to spin us up a Singapore and UK colo servers.

If you spent $X on a SAN rack 10 years ago, that same $X now buys you a fully-managed rack with NVMe drives and redundant, hot-swappable everything. Your Ops staff will get an email about a drive or hardware failure and the tracking number for the replacement part that’s already left the depot and can be replaced at their leisure.

You experienced badly run shop. You can have remote hands (if you don't have your own DC, but buy colocation service). You have spare disks in server. I don't know, how end-to-end encryption is linked to bare metal, those physical servers are several abstraction levels below.

I think the implication was that Amazon provides more than just bare metal to lock you in.

A startup worrying about long term cloud implications is like a hot dog vendor in new York worrying about how his branding will be received in India when his hot dog empire expands there. 99.99% of startups will never have to worry about migrating off cloud, and if they do, they will have succeeded beyond their wildest dreams, and will be once sequential. Nobody is losing sleep over the prospect that their startup will only be worth 22.4 billion in 17 years, but if they eschewed the cloud and built in efficiencies from the get go the company would be worth 23.7 billion. Not to mention the fact that worrying about long term things like that would take focus off the present and in tease chances of failure.

Nitpick, but we don't hear about the companies that failed in the mid-stage because they ended up being consumed by their AWS fees.

>This seems like a well known trap that people would know how to avoid with a little planning, but I guess they have no one thinking that far ahead

But you see part of the magic trick was Amazon spent years having companies like Garnter proclaim that cloud was cheaper, and nobody could possibly do on-premises IT cheaper because of economies of scale. As a result you've got CIOs everywhere trying to make a name for themselves by driving their company to the cloud to show immense cost savings. They don't have time to be bothered by the actual financial models or real costs. By the time finance realizes what happened they'll be on to the next gig.

It was honestly brilliant, the number of "cloud first" strategies that originated in board rooms filled with people that don't know the first thing about IT is kind of disgusting.

It can be a "trap" but also the best decision for a startup. It probably is actually.

When Companies Should Go to the Public Cloud + When they Should Return

https://blog.min.io/repatriation_wave/

DropBox did it [0].

[0] https://techcrunch.com/2017/09/15/why-dropbox-decided-to-dro...

"Why Dropbox decided to drop AWS and build its own infrastructure and network"

tl;dr: to save money (and to fix issues and irritations with AWS)

And then moved back: https://aws.amazon.com/solutions/case-studies/dropbox-s3/

I am sure they could negotiate excellent conditions because of their previous setup.

For analytics. I don't think the entire thing runs on AWS anymore.

Every truly large corp should have inhouse options for negotiations with AWS.

And stopped innovating at around that time

Twitter was on a laptop, then it was in ruby, then it was in java. Companies grow and change, and if your seriously scaling, then your behavior in different environments needs to adjust. It's simply a strategic planning issue for the CIO.

If moving to another cloud provider, or on-prem, has a "very high cost" then you're doing it wrong.

I don't think you ever worked on a large enough application to appreciate this complexity.

Have you been part of any large migration? Even if you try to stay “cloud agnostic” (no using Terraform doesn’t count. Each provisioner is cloud specific), everything takes time at scale. Not to mention data migration costs, project management, regression testing, a company of any scale is probably still partially on prem and has all sorts of interconnecting network points with the cloud provider, IAM, etc.

Of course any migration has costs. But if they are higher because of deep entanglements in specific cloud vendor services, that is a vulnerability. It's already demonstrated that the big cloud providers will cut you off with little warning if you run afoul of their political sensibilities. So the only way to use those services is in the most generic way you can, so that migrations don't cost more because of it.

And if you are using cloud just as an overpriced colo, you aren’t really getting any advantage of being in the cloud. You’re going up spend more money hosting all your own infrastructure on VMs, more time on maintenance, and does it add any business value?

Do you not use IAM? Cloud specific infrastructure as code (again TF has cloud specific provisioners), permissions, compliance, training.

The articles mentions many basic advantages of using cloud platforms as simple infrastructure, without entangling your usage with vendor-specific, proprietary services. Scalability is an obvious one.

I’ve seen this parroted so often with anything infrastructure related. Where do you magical people work that your company has managed to somehow find the time to build out infrastructure that you can swap out to another provider on a whim? My assumption is you must already self-host and run everything without using any external providers.

> somehow find the time to build out infrastructure that you can swap out to another provider on a whim?

You can very much do that while staying in the cloud. It's all about engineering for flexibility.

Ok.. so that still doesn't answer my question. How do you get the time and resources allocated to make your entire platform completely agnostic of specific cloud services while still operating in the cloud? Can you name a single company where that's done and they talked about it at all? We're talking about at least tripling developer workloads and general operating costs here and finding developers who can navigate different cloud and on-premise providers for... what?

Even small companies have multiple engineering teams, and each of those teams has particular needs, and each of the engineers, analysts or general users may rely on some service AWS, GCP or some other cloud provider has. Any single person can now break this ideal flexibility requirement if it makes their life easier.

Happens all the time. I personally worked at co that went from aws to gcp and now are moving to azure and know of a few who did the same. So if you can change clouds you could move to onprem presumably if you can find folks who know how and don’t already work for your cloud provider

My question was towards:

> It's already demonstrated that the big cloud providers will cut you off with little warning if you run afoul of their political sensibilities. So the only way to use those services is in the most generic way you can, so that migrations don't cost more because of it.

You are stating that you moved from one cloud to another. The question is how you manage to build a company of any size that can just flip a switch and move to a new service provider without needing to transition teams to using new services and that costing money. If we're talking about moving a simple server from AWS to GCP, sure that's no problem. But if you or anyone at your company is actually using a cloud providers services for more than hosting a server, there is no possible way you can just flip a switch and migrate over.

I’m talking about annual cloud spends in tens of millions here. It’s doable but you have to design for this or it takes a long time

And if you are spending in the tens of millions, any cloud provider is going to negotiate discounts with you to keep you on board.

You’re not getting retail price.

So you’re still spending more trying to maintain “cloud agnosticism” by hosting everything on VMs instead of using managed services.

The discount is not as much as you’d think. They’ll do it for some high margin services like storage and gpu (where markup is just ridiculous) but won’t budge on network for some reason. In the end they gon getcha anyway

Oh and btw the models that i had worked out had ~10-12mo breakeven for colo vs cloud including the discount since you’re doing 3y commit anyway. If you’re going with private dc you have to be really large to make it work but if you have large egress charge it becomes instantly worth it (like in dropbox case)

How’s that working for Dropbox? They aren’t exactly a shining example of moving from the cloud and becoming extremely profitable. They still haven’t made a GAAP profit.

Like Jobs said, Dropbox is a feature not a product.

For the same price you pay for Dropbox and 2TB, you can get the full Office 365 suite plus 6TB of storage or the entire GSuite.

And do you have multiple redundant DCs? Are all of your database writes duplicated 6x across three DCs like Aurora?

> How’s that working for Dropbox? They aren’t exactly a shining example of moving from the cloud and becoming extremely profitable. They still haven’t made a GAAP profit.

That’s just silly - dropbox is profitable if you exclude stock comp unlike box inc (who afaik runs in cloud) and also has higher margins (big shocker here). From techincal side i hear it worked out really well for them too.

> Like Jobs said, Dropbox is a feature not a product.

What does that have to do with anything? Not helping your argument at all

> For the same price you pay for Dropbox and 2TB, you can get the full Office 365 suite plus 6TB of storage or the entire GSuite.

Should be painfully obvious to anyone arguing in good faith why that is. Also do you by any chance think google runs gsuite on aws or gcp?

> And do you have multiple redundant DCs? Are all of your database writes duplicated 6x across three DCs like Aurora?

Yes, you run mutiple redunant colos/dcs and peer them. You’re making it sound like rocket science but it really isn’t. If Internet Archive can do it with their resources (not a dig towards their talent at all) so can unicorns flush with vc cash who supposedly hire “the best”

Dropbox is profitable as long as you exclude the fact that they have to pay people…

How many people who are comparing their costs to using a cloud provider are ignoring redundancy?

And I’m sure your average startup has the capability to maintain their own redundant data center. Besides, does it add business value? Would they really be better off hiring multiple infrastructure people or just paying a cloud provider.

It doesn’t matter why it’s the case. People keep bringing up Dropbox as a shining example of moving away from a cloud provider - and still not being profitable. But ignore a little company - Netflix - that did just the opposite.

It seems that the question you are answering, and that I don’t disagree with and am not asking is ‘can you move to a different service provider?’. Not ‘how do you move to a different service provider without incurring costs to transition to different services?’, which is what my question actually is.

The cost is one time and can be relatively contained. Also if you design for redundancy you dont have to do a hard cutoff - just deploy your next region to a dc/different cloud and then gradually drain old ones.

The difference is the bottleneck in skilled cloud practitioners. There aren't that many people who can deploy these systems at scale, and they are best rewarded at the centre of the system. So the system centralizes. Enterprises would love to duplicate AWS, but they don't have the ability to each hire similar calibre talent and keep them engaged on their pet clouds.

I disagree with this characterization as incomplete. In the post-SUN era there were plenty of people who could build and administer a dozen forms of server OS+hardware, but a confluence of factors caused a greater tide to move greater numbers, in another way.

Some commerce did need to scale fast and easily, and AWS provided that with great strength. But I will add that the management and executive branch were constantly and relentlessly seeking to undermine the bargaining power, day-to-day operations authority, and pay of skilled engineers who could do that kind of setup. Not that it was easy or risk-free .. I did see some powerfully strained eyes and nerves on those individuals who built and ran those kind of servers at various times. But do not leave out the manipulative and gaming motives of executive management to dumb-down admin requirements, seize authority via legal means, and seek the lowest wages across borders while doing so ..

> This is the best-executed vendor lock-in I know.

Another good one is microsoft initially tying excel, word, access, etc to windows. And then IIS, SQL Server, etc to NT. An entire generation of businesses and managers got tied to windows this way.

In a way facebook and social media is a form of "vendor lock-in". The network/community itself and your "investment/time" in facebook makes it costly for many to move to another "facebook" or give up facebook altogether. The more time you spend on facebook and the more people you bring onto facebook, the more "locked-in" you are. It's quite ingenious and I suspect facebook understood this very early on.

The former is lock-in. The later example is better described as "network effects", i.e. Facebook is compelling to you because all your friends are on there, and the more of your group that engages on the platform the more valuable it is to you and the harder it is to get everyone to jump in mass to another platform.

Having taken a large organization's entire stack the other way in 2016 - from dedicated metal colocation to AWS, after a series of devastating ddos attacks on our datacenter - I've remained cautious and judicious in how I use cloud services, to attempt to prevent lock-in, for this reason. I build all critical systems (including outbound mail - which requires negotiating with Amazon) on raw EC2 VMs, allocated drives and RDS images, and more or less balance workloads between them manually. This involves a bit more hassle than e.g. putting frontends on elastic beanstalk with automatic load balancing, or using buckets or any of the nicer cloud features... I basically treat AWS as a colocation facility where I don't need to get anyone on the phone to reboot a machine at 3am. I've been told I'm "misusing" or even "misunderstanding the purpose of" the wonderful services provided by AWS. But for me, the upshot is that I could probably port the entire infrastructure to any other server farm within a few hours if necessary.

I like how some of the services are very cheap (S3, Lambda), but then at some point you end up using a bunch of other complements, one of which is very expensive and makes the profit.

At which point, as you say, the cost and complexity of a move is huge so you are stuck with a recurring bill.

Another interesting thing is that you pay for 100% of a VM virtual CPU even if it is only utilised 40%, but the cloud is selling your other 60% capacity to another user.

They have no incentive to charge you “based on CPU usage” for VM’s that are always on, even though it is possible. I think this would be a better approach than Lambda - to dynamically add more cores to your OS as you need them.

> Another interesting thing is that you pay for 100% of a VM virtual CPU even if it is only utilised 40%, but the cloud is selling your other 60% capacity to another user

AWS don't overprovision CPUs outside of the t instance range, in which you get vCPU credits, not full vCPUs. If you rent an apartment but don't use it, is it the landlord's fault you don't? The landlord can't rent the same appartement to another client.

Especially on AWS/GCP/Azure, where there are many alternatives ( Lambda, Google Cloud Run, Google App Engine, etc.) where you pay per compute used.

At the hyper visor level, I was under the impression that many VM’s are scheduled onto the same hardware CPU core.

One hardware core can be running 5 VMs at 20% capacity, and each VM owner will pay for 100% of their CPU capacity right? So in this case the landlord is renting your unused apartment to other customers then evicting them faster than you can open the door.

My point is cloud is benefiting from that ability and not passing it on the customers, which they could.

Your impression is correct for VPS-style "clouds", like Digital Ocean and co.

Its fully wrong for AWS, where a vCPU you rent == physical core/thread.

How do you know this implementation detail?

That statement is counter to how "vCPUs" related to "CPU" since early cloud implementation. AWS not taking advantage of the much more efficient "oversubscription ratio" versus thread pinning - unlikely.

Are you certain of this statement?

The solution is an internal marketplace. Large enterprises formed from conglomeration often flub this. They often end up with a half dozen or more datacenters and ops groups. What do they do?

Consolidate them, and a knife fight happens between middle management machiavellis and the one that wins is not necessarily the best on price or features. It's the one who is best at winning reporting hierarchy wars.

Instead they should pit these groups against each other as internal options for hosting to various services that provide what is needed for growth, adaptation, or cost.

AWS should simply be one of those options. Really it can be the prototyping phase, and then budget squeezing will move the app to other (hopefully cheaper) inhouse options.

What is frustrating from my years in the big corps is that AWS was cheaper than the internal "clouds" or other relationships.

The paradox is that in spite of the huge costs, very large companies are still moving to (or using) the cloud.

I've had this discussion many times where I work, and we've refrained from using cloud services since they are extraordinary expensive compared to what we can do internally.

Yet, people seem to be convinced that prices are ok, and that going outside will let them reduce their infrastructure footprint.

For some reason I don't quite understand, people tend to be attracted by the cloud, even when it makes no sense economically whatsoever, and talking sense into them is surprisingly difficult.

So yes, it seems obvious that people will choose the most cost effective solution. However, it seems like they don't!

>The paradox is that in spite of the huge costs, very large companies are still moving to (or using) the cloud.

In my experience when a large corporation moves to the cloud it has less to do with pricing and more to do flexibility. It's easier to get a budget from IT and do whatever you want in cloud, then to have to wait weeks/months putting in a PO for hardware, getting access to those machines and installing what you want.

This is true. Where I work a PO was issued several weeks ago for a rack of servers, and delivery will be in about 3 months. If cost were no object, those resouces could be provisioned with a cloud provider today. And getting from requirements to RFQ to PO took another 3-4 months.

People are attracted to it because owning stuff sucks. Renting stuff is more expensive but sucks less. This is why businesses often prefer opex to capex even if it is pretty significantly more expensive. It isn't just an IT phenomenon.

Yep people will pay through the nose to move from capex to opex. It baffled me at first, I think it baffles me less (but still a bit) now. Maybe one day I'll understand.

What about housing? Why do people still buy housing?

Two reasons: 1. For the investment in the land the house sits on (the house itself is a depreciating asset, but the land may be a good investment), and 2. For the ability to control their own destiny, to be able to make improvements and to have a place to live that does not rely on a landlord's willingness to continue signing leases with you. Note that #2 is totally analogous to owning your own server infrastructure. You definitely give up control when you rent rather than buy compute. But I think people are more sensitive to this in their personal lives than businesses (especially small businesses); after all, you're stuck with your personal life, you can't sell it or declare bankruptcy or just shut it down if things don't work out.

The list of things that I don't understand only seems to grow longer, so I'm not sure, sorry.

It's an accounting phenomenon: Deducting expenses reduces income tax, which is levied on net income. It is also beneficial when considering the time value of money – money available at the present time is worth more than in the future due to its earning capacity.

Some companies have terrible IT outsourcing and procurement practices such that getting something comparable to aws infrastructure is effectively not possible, as it would require massive internal change whereas moving to the cloud is easy to get buy in.

Also regular devs don't realize but frequently internal IT departments charge obscene amounts of money for managed services.

Devs look at Digital Ocean or Hetzner VM costing $10 for a ton of RAM, storage and bandwidth when the same thing internally in a bank or other big enterprise can cost $100-150. AND be delivered in 3 months, if it's ever delivered.

Absolutely, bargaining for cheap service fees from these suppliers, only to be charged greatly for simple hardware and anything outside the standard list.

> Yet, people seem to be convinced that prices are ok, and that going outside will let them reduce their infrastructure footprint.

That's where you get it all wrong. It is not about the price, at all. It is all about avoiding large upfront capex with a small periodical opex, and in the process have virtually boundless growth potential.

Think about it: when you happen to need a bit more computational resources to run an app, is it easier to convince your boss the need to pay, say, 100€/month for extra VMs, or is it easier to convince your boss to shelve $2k to buy an extra rack? And how many times are you willing to have that same conversation with your boss whenever you run out of computational resources?

Aren’t the real figures more like $200 vs. $700? Cloud stuff only really makes sense when your resource requirements as super spiky.

100% agree on shifting capex to opex though, capital works teams are super in favour of that, oddly enough.

> Aren’t the real figures more like $200 vs. $700?

No, not really. You can hardly find old used servers for less than 1000$, and you need to use beyond a a1.metal instance with 16vCPU and 32GB of RAM to even come close to that cost

> Cloud stuff only really makes sense when your resource requirements as super spiky.

No, not really. Cloud stuff makes all the sense in the world if all you need is a VM connected to the internet working 24h/day without problems. You can get that for less than 5$ a month, which would require 3 or 4 years to pay back if instead you waste your money on hardware.

Not my experience but ymmv (and clearly has).

> You can hardly find old used servers for less than 1000$

I picked up a second hand Dell PowerEdge 610 with 96GB of RAM for $400 AU, it's overkill by a large factor for what I need but damn. Either I was super lucky or servers aren't (weren't?) expensive.

> You can get that for less than 5$ a month pocket change and host it

Until you actually use it and then it ramps up rapidly to the point where you'd be far better off just buying a server.

Good luck getting spares, patched BIOS/firmware, patched DRAC, or anything else resembling support for that old Dell in a timely fashion.

And good luck selling to enterprises when the vulnerability scanner lights up like a Christmas tree with critical and un-patchable flaws on your switches/routers/UPS/PDUs/storage arrays/servers/door locks/cooling system/fire suppression/cameras. Double for the DR site. Because you need all that stuff, and you have to maintain it all 24x7 which means lots of salaries.

The same reason big companies pay premium for external consultants instead of directly hiring. You don't want to deal with employees, as well as you don't want to deal with HW, data center, and all that comes with it. You know you are paying premium, but you are also delegating tons of headache and responsibilities.

The mining companies I work with often hire equipment semi-permanently instead of buying it. It’s an insane cash crop for the hire companies because the payback time on the equipment is a few months and then it’s pure profit, but apparently a $700/week car rental payment is easier to get through accounts than a one off $50k purchase.

Well I guess by renting they can get the latest and greatest machine available instead of dealing with an old machine bought say 5 years prior. If the machine breaks, well you just get another one. No maintenance cost for the company. I mean it depends. Buy it is not always the best thing to do. Renting may be more convenient sometime.

You'd hope so but these are the same machines on site that have been there for years and they're still paying the same per week for them. It literally just comes down to the capex/opex dilemma, individual caps on expenditure without review, and lack of motivation to improve efficiency if it doesn't land in their pocket from everyone else.

Running in house, you need to know what you want, and you need to be pretty accurate in your estimate. And you need to pay today for what you're going to need in six months, a year, three years.

And if you're wrong about those estimates, the costs are real. Overinvested in hardware upfront and your project will never break even. Underinvest, and when the marketing campaign kicks in and you get flattened on cybermonday, there is no limit to how much revenue you might miss out on.

I'm willing to pay more money to reduce the downside of my being wrong. Cloud services offer so much flexibility that I don't even have to make guesses about some things beyond my information horizon.

On prem might be cheaper but it is far from an equivalent service.

You don’t need to be particularly accurate. The cloud is so expensive you have plenty of wiggle room while saving money. Worst case you can still use that hardware on another project. Going the other way adding more hardware 6 months out is generally trivial.

> Am I alone in not seeing a paradox here? It's no different from a variety of things that make sense at some scales and not at others.

Yes, it's not a paradox at all. Use this thing until it doesn't make sense, and then don't use it anymore (or be comfortable with the lack of optimization). This is not a paradox.

I assume the article gets more attention if it uses the word paradox.

A paradox is only at the most coarse-grained level. Cloud makes sense because of its cost-efficiency in scaling from nothing upwards. Then fails because of cost involved in keeping it running at scale. The thing to recognize is that what cloud is good for is rapid change. Once you have something stable, even at not-huge scales, bare-metal hosting may still make more sense.

All this happens willingly unless growth happened with your eyes closed, or really had no choice keeping up with it. Sure, use the cloud, but don't use every cloud-proprietary convenience. Use opensource software in a near-standard way. Even if using the cloud-vendor's offering, refrain from using proprietary extensions. If your plan is to grow and get out before things stabilize, then that's on who's come onboard, not resisted the conveniences, and remaining.

The paradox is the sunken cost of your product easily becoming coupled to the cloud.

Moving office is relatively simple, you call a moving company to lift all your chairs, print new business cards, hire a bigger maintenance crew and it's done within a week.

Moving cloud could in worst case mean a 6 month complete rewrite of your core product depending on how much dependent you are to CosmosDB, Lambda or other shiny AWS stuff.

You could say buying a machine you previously rented is the same thing but when looking at code migrations, rewrites and launching new infrastructure cost time from R&D, not just money.

I think point guy tried to make is, that ton of big companies, that would/could benefit from in-house solution still go for full cloud based one without thinking about alternatives.

Reply on news.ycombinator.com