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

Dropbox saved $75M over two years by building its own infrastructure (2018)

geekwire.com

Post reply on HN
Print view View original

201–210 of 218 posts

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#201

Post by georgyo »

peterwwillis wrote: post #198
calderarrow wrote: post #3

I think this quote really hits the nail on the head and confirms what a lot of people may have intuitively known about the value of cloud providers: But once certain startups turn into big companies with hundreds of millions of users, with computing needs that they’ve come to intimately understand, it can be far more efficient to set up computing infrastructure designed exactly with those needs in mind. I think the m…

Why do people buy coffee from Starbucks? Is it because Starbucks has the most efficient cost outlay for long term investment in a user's needs? Or is it because people just want some damn coffee and there's one on every corner? Or is it because, in a world full of places to buy coffee, one place gives you everything you could ever dream of in a coffee place ? That is what the big cloud providers are. They are Starbuc…

This is not a great analogy. Sure people go to Starbucks for one off coffees and sure some do that exclusively. But almost every startup to enterprise company has an expresso machine in the kitchen.

So in most cases, if I want a coffee I don't go to Starbucks. They I go to the kitchen, it's faster, cheaper, and more convenient than going to Starbucks.

However, if I'm out and about it doesn't make sense for me to invest in temporary coffee infrastructure. In those cases it's easier to go to a cafe like Starbucks.

This matches the idea of the article. If you have consistent demand it makes sense to buy infrastructure to meet that demand. But if the project is temporary or extremely bursty it may make sense to have someone else do it for you.

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#202

Post by lazylizard »

we only have less than 60 racks. given every 2u can fit 4 x rtx and 4u can fit nearly 1pb raw. in general its cheaper than what people can rent to us for.

also. since we already exist. we run mail, ldap n a few other bits as well. i still think its all cheaper than 3rd party hosted.

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#203

Post by fakedang »

jedberg wrote: post #152
dec0dedab0de wrote: post #132

Earlier quoted context omitted.

Sure, but if you are using the taxi often enough it may be cheaper to higher a full time driver.

Assuming you are qualified to hire drivers. And know where to find a good one. This is what people forget with the cloud. That you're paying for their expertise.

Or you could hire off them?

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#204

Post by fakedang »

ralph84 wrote: post #48
johnklos wrote: post #29

"Cloud" handles the problem of people who buy based on who takes them out to lunch, buy based on wanting to have bragging rights, buy based on the idea that "nobody ever got fired for buying Dell / HP / whatever". Sure, you can buy and colocate a dual processor Xeon Platinum bought from Dell, but when you can get significantly more than twice the performance per dollar buying an Epyc system, it's stupid to do it. Sto…

What makes you think Amazon and Microsoft don't take customers out to lunch (pre-covid)?

Lol exactly. Amazon and Microsoft always take potential customers out for lunch (or dinner). My fiancee and I have attended a number of "Azure" dinners, where Microsoft funded our expenses. The only qualm was that we had to sit through lectures where salespersons would go on about how Xamarin was the next hot shit, which bored her, a medical doctor, to death, and managed a few chuckles from me.

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#205

Post by pjmlp »

ericbarrett wrote: post #61
calderarrow wrote: post #3

I think this quote really hits the nail on the head and confirms what a lot of people may have intuitively known about the value of cloud providers: But once certain startups turn into big companies with hundreds of millions of users, with computing needs that they’ve come to intimately understand, it can be far more efficient to set up computing infrastructure designed exactly with those needs in mind. I think the m…

You’ve described the historical value of cloud computing perfectly. That said, I think the days where all but the largest or most stubborn companies run their own datacenters are coming to a close. The problem will be finding skilled labor. Short-haul networking, power configurations, thermal load, hardware maintenance; these and many more are specific skills that can’t be learned overnight. Data center work used to…

If you think you can manage cloud deployments without skilled labour with cloud expertise, you are in for a surprise.

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#206

Post by dddw »

Aperocky wrote: post #45
holoduke wrote: post #43

We are a startup. We decided to host our full stack in self managed vpses running debian. We started at one provider, but because we have everything (except the hardware) under our own super vision, we were able to move to another provider (hetzner) in just a few days. Moving away to another provider wouldn't be an issue. I believe the key is to not use any of the propitiatory services offered by all these big player…

What are the size of the servers?

Yeah 1000 bucks pm for 15 servers... should be quite decently specced

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#207

Post by einpoklum »

spullara wrote: post #189
einpoklum wrote: post #160

Earlier quoted context omitted.

Elasticity? Fine. So their, say, single rack will sometime have limited load and be under-utilized. About the up-front investment - most hi-tech companies are a massive initial up-front (or nearly-up-front) investment.

I was talking about at scale, not a rack. If you can get by with a rack, you will pay more for the people to support it than the incremental cost of the cloud.

> If you can get by with a rack, you will pay more for the people to support it than the incremental cost of the cloud.

Probably a whole lot less.

At larger scale - I would guess it's the same thing. If an organization needs more than a rack during peak use, it can probably benefit from setting up its own infrastructure. Only in the uncommon case of short extreme peak use and almost no use most of the time does such elasticity make a could solution attractive. IMHO.

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#208

Post by evrydayhustling »

smarx007 wrote: post #123
evrydayhustling wrote: post #105

Earlier quoted context omitted.

Check out these curves: https://www.stayclassyinternet.com/articles/investigating-AW... Amazon competes with other cloud providers and with build-your-own continually, and use price segmentation to stay on the right side of the equation for as many people as possible, while still skimming as much premium for themselves from each user group. Historically, that’s meant big price drops in most categories that would make…

Yes but no. The figure you reference shows the lowest price of a lowers tier available. I am pretty sure if you cannot switch from S3 with full availability to S3 Glacier, your savings will not be SO impressive as that fall-off-the-cliff figure.

The figure shows full-availability S3 at multiple price tiers depending on volume. The article discusses that price reductions were biggest for high volume -- in other words, AWS reduced price for exactly the customers who are most equipped to build-not-buy. Smaller use-cases won't see prices go down as much -- but they also won't get paid back as quickly for rolling their own.

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#209

Post by thor24 »

easton wrote: post #167
thor24 wrote: post #118

Earlier quoted context omitted.

Instant Elasticity in Cloud is a myth. If you think you are going to get 1k hosts just like that from AWS you will have an unpleasant experience. I work at one of the decent size tech company and we are split between cloud and on prem. From our experience you have to inform AWS/GCP in advance (sometime way early) if you are looking to meaningfully increase capacity in zone/region. Sure, auto scaling few hundreds of h…

How early is "way early"? Because as long as it's shorter than the two-three weeks it'd take to order boxes, rack them, provision them (which would be automated but might still take a afternoon), deal with any QA hiccups... I'd much rather call my AWS rep and say "can we add 30% by Thursday" and have them figure it out (and at such a large scale you might be able to spread it out across a couple regions anyway unless…

From what I have seen it is actually of the same order or sometimes more. In one of the region/zone we add few hundreds hosts every week but that is after telling them we plan to upscale this in this region upto some big X number.

Re: Dropbox saved $75M over two years by building its own infrastructure (2018)

#210

Post by thor24 »

echelon wrote: post #119
thor24 wrote: post #118

Earlier quoted context omitted.

Instant Elasticity in Cloud is a myth. If you think you are going to get 1k hosts just like that from AWS you will have an unpleasant experience. I work at one of the decent size tech company and we are split between cloud and on prem. From our experience you have to inform AWS/GCP in advance (sometime way early) if you are looking to meaningfully increase capacity in zone/region. Sure, auto scaling few hundreds of h…

Why not go fully on-prem then? You can run kubernetes locally. Are managed data stores that attractive? You can pay for on-prem management. What workloads are in the cloud versus on-prem?

Right now there is no specific distinction between what we want to run in Cloud vs On Prem. Important thing to note here is we use Cloud as an IaaS only. We have our own stack which sort of prepares the hosts before it is ingested into clusters as usable capacity.

We actually recommend not using custom cloud providers Databases or any other value added services.

Why not completely either way (on prem vs cloud) is something that happened way before I joined the group but I think the main reason is to have a tactical edge in the long run such that we avoid lock in. I guess in some ways it helps us negotiate pricing better.

Imagine moving a certain workload from GCP region to an AWS region as part of a failover drill.

Post reply on HN