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

171–180 of 218 posts

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

#171

Post by ebiester »

mike_d wrote: post #154
ebiester wrote: post #6

Just as a reminder: You are not Dropbox. You personally need to run the numbers with both scenarios. Depending on your use case, it may be cheaper to be in AWS, and it may be cheaper to be co-located. If you are as big as Dropbox, it may be cheaper to build your own data center. It reminds me of the arguments for and against K8s. Most of the discussion is based on use case, and not considering the solution space for…

> ...and it may be cheaper to be co-located. If you are as big as Dropbox, it may be cheaper to build your own data center. Nothing I have seen indicates dropbox "built its own datacenter." It looks like they are in Equinix, CoreSite, and DRT colocation.

You're right - I misspoke, and that likely makes more sense at their scale. At some point, I am sure they will again run the numbers.

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

#172

Post by ericbarrett »

Shorel wrote: post #113
ericbarrett wrote: post #61

Earlier quoted context omitted.

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…

You say "stubborn" but when the savings can be in the millions of dollars, it becomes logical to own a datacenter. I can't believe datacenter people suddenly all disappeared or are retired. If the pay scale is down, they will surely appreciate working for a big company with datacenter needs. Obviously, this only applies to huge companies like Dropbox. Everyone else is better served by AWS.

This is explicitly what I’m calling out at the end. They didn’t all disappear or retire; it’s an ongoing process. Think COBOL programmers.

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

#173

Post by mxuribe »

jascii wrote: post #31

I often wondered why we aren't seeing more hybrid solutions, base systems running on minimal owned hardware with cloud options for peak-load and DR. I guess part of the problem with that is the cost of maintaining different platforms, but there are ways around that.

I can see that more for mid- or large-sized enterprises...If i was such an org, and already had some investment in infrastructure, then hybrid is what i would look at first. Not sure that i hear that so often though. Mostly, what i hear about is, "let's go to the cloud!". Oh well.

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

#174

Post by jonfromsf »

Shorel wrote: post #113
ericbarrett wrote: post #61

Earlier quoted context omitted.

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…

You say "stubborn" but when the savings can be in the millions of dollars, it becomes logical to own a datacenter. I can't believe datacenter people suddenly all disappeared or are retired. If the pay scale is down, they will surely appreciate working for a big company with datacenter needs. Obviously, this only applies to huge companies like Dropbox. Everyone else is better served by AWS.

You can rent racks of servers from a dedicated hosting provider and spend 10X less than AWS. You don't need to do you own wiring and HVAC and shit.

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

#175

Post by phil21 »

mikepurvis wrote: post #21
rossmohax wrote: post #9

I remember Gitlab blogged that they are going to use baremetal instead of Azure with some numbers showing benefit of doing so. AFAIK they quickly reversed decision and didn't even try to implement that plan, moved to GCP instead.

I'm not privy to any particulars specific to GitLab's situation, but it's possible that a little sabre-rattling caused a better deal to appear which changed the calculus on migration.

My limited visibility take is that their board (or someone higher up, my memory is fuzzy) told them to use the cloud, and that was that.

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

#176

Post by oblio »

mbesto wrote: post #60
H8crilA wrote: post #24

So they saved next to nothing? Their yearly operating expenses are in the order of $1.3B, which doesn't even include cost of revenue. $75M savings versus $2600M operating expenses over those 2 years comes out to about 2.8% savings.

> $75M savings versus $2600M operating expenses over those 2 years comes out to about 2.8% savings 2.8% of a big number is still a bigger number than 2.8% of a small number. Not sure what point you're trying to make here...

Opportunity cost plus the risks of this migration.

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

#177

Post by cogman10 »

heipei wrote: post #92
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…

The thing is that if you don't use any of the proprietary services offered by the bigger players they look at lot less appealing. AWS with just EC2 is one expensive-ass server provider with murderous rates for egress traffic.

Really depends on what you are trying to do.

I work in a company that is currently trying to move over to AWS from on prem servers.

Why? Not for AWS proprietary software. That's completely an afterthought (In fact, the first round is likely going to be those expensive ass EC2 instances).

The real reason we are looking at moving to AWS is because they have data centers across the globe and we don't. So even though EC2 is expensive, it's far less expensive than trying to navigate the waters of setting up a datacenter in France, china, and India.

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

#178

Post by ehsankia »

dilyevsky wrote: post #95
benhoyt wrote: post #82

Earlier quoted context omitted.

Not a bad analogy, but the taxi also comes with a driver, which makes it expensive but allows you to get work done in the back while you're en route. It'd be like a cloud provider that came with a devops person dedicated to you.

Except when driver decides to take couple days off and you can’t get anywhere or their cars breaks down for a day and they say it’s still within slo because the radio still worked (looking at you GCP)

Except in this analogy, a taxi breaking can be quickly replaced with another taxi, the issue is when the whole taxi network goes down which is more rare. If you had your own car, if the machine went down, it would be a lot more work to fix it yourself.

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

#179

Post by ehsankia »

whimsicalism wrote: post #46
H8crilA wrote: post #24

So they saved next to nothing? Their yearly operating expenses are in the order of $1.3B, which doesn't even include cost of revenue. $75M savings versus $2600M operating expenses over those 2 years comes out to about 2.8% savings.

> So they saved next to nothing? Their yearly operating expenses are in the order of $1.3B, which doesn't even include cost of revenue. Wow thank god some of y'all aren't running these businesses. It still surprises me every time I see failures to think marginally.

That's assuming the numbers shown fully take into account all the tangential cost related to the transition, and not just "cost before vs cost after". If you spent 500M in developer hours doing the migration, that just wiped like 10 years worth of savings. Also it's not clear if in 2 years you really get to see the cost of maintenance and breakages. One big breakage could wipe off that 75M very quickly. Of course Cloud also has its fair share of down time, so you'd have to compare longer term to get a better sense.

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

#180

Post by mbesto »

H8crilA wrote: post #96
whimsicalism wrote: post #46

Earlier quoted context omitted.

> So they saved next to nothing? Their yearly operating expenses are in the order of $1.3B, which doesn't even include cost of revenue. Wow thank god some of y'all aren't running these businesses. It still surprises me every time I see failures to think marginally.

~3% boost to margins is very welcome for a cloud company (would be amazing for a retailer). But what about the risks? What if you need to scale down the operation? What if your hardware team ends up costing more? In particular, the R&D department (cloud providers depreciate hardware over 2-3 years, which matches the progress of technology and the speed at which things become obsolete). $35M per year only pays about 5…

> One very good strategic reason to do this, I think, would be to gain independence from cloud providers, who unfortunately happen to be your major competitors.

All of the questions you asked before (risks, capex, etc.) all would have been answered during an assessment period. It's almost as if you assumed that none of that was done and that their sys admin group just through together a proposal to move to a colo and their executive management made a decision to do so. That's not how this works.

The number one priority of this type of initiative was cost and reliability. Going exclusively cloud is not a silver bullet for every company - that trope needs to die.

Post reply on HN