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

141–150 of 218 posts

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

#141

Post by chewmieser »

sharadov wrote: post #8

I work for a company that does about 100M revenue per year. Run everything on prem, we did the numbers if we moved to AWS, it was almost 10 X of our current spend. Again , depends on your business, ours is complex and compute intensive. But at scale, on-prem always wins.

I work for a 100M rev/year company as well and we used to run our ecommerce site on-prem as well. We knew of the increase to cost for moving our infrastructure to AWS but we're happy to bite that bullet for the vast array of services we now leverage in AWS as well as faster site response times and increased reliability.

Just saying spend isn't the only factor here.

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

#142

Post by chris11 »

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…

> it should be easier to migrate off these platforms, since their pricing is optimized for users of all business sizes and use cases

Isn't the financial implications of capex vs opex a huge consideration? I've heard that opex is is a lot simpler to account for. Technically once you get big enough cloud becomes more expensive. But hiring people to manage both your own datacenters and cloud services does complicate things. I find it understandable that companies are willing to pay more for cloud providers if their core business doesn't require expertise in cloud computing.

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

#143

Post by data_spy »

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…

Agreed. I work at a company that does 150 Billion transactions a day and it is cheaper being in the cloud than on prem for us. Maybe the variable cost is more but the fixed cost is a lot less for the business.

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

#144

Post by ryanmarsh »

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…

all but the largest or most stubborn companies run their own datacenters are coming to a close

You're right about the stubborn part, but not necessarily largest. Many IT leaders at companies of all sizes have their political capital tied up in the data center. What you're talking about doing (with cloud adoption) is outsourcing 90% of what they control. For them it's existential, uptime and agility be damned.

Serverless is even worse for them, as far as IT fiefdoms are concerned.

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

#145

Post by dastbe »

justincormack wrote: post #104
HatchedLake721 wrote: post #98

Earlier quoted context omitted.

> ... and not necessarily all that frequently. Did you miss where AWS reduces prices multiple times a year? I don't know the latest figure, but as of 2018, they reduced prices 67 times since launch in 2006.

There have been 6 price reductions this year https://aws.amazon.com/blogs/aws/category/price-reduction/ you are probably not using any of the services in question. The only S3 price reduction I can find was in 2012 https://aws.amazon.com/blogs/aws/amazon-s3-price-reduction/

(i work at aws)

fwiw, a quick google search for `s3 price reduction` yields one from 2016: https://aws.amazon.com/blogs/aws/aws-storage-update-s3-glaci...

there's also one-zone s3 storage options which are cheaper as well

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

#146

Post by mrlala »

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.

Even if you want to argue they saved next to nothing (which is silly to say, $75m is $75m).. they also have CONTROL over what they are doing and are not dependent on other services. That is a huge advantage.

So even if it were a wash financially, it's still a huge plus.

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

#147

Post by birdyrooster »

ericbarrett wrote: post #121
birdyrooster wrote: post #85

Earlier quoted context omitted.

They aren’t shrinking at all, we have the same talent pool and are trying to thinly spread it over all of the problems we can solve with networked computers. The amount of problems you can solve with networked computers has scaled way faster than our industries capability to find talent and train employees.

Who is “we”? Always curious what remains of my field—I’ve worked in the storage and infrastructure industry for over 20 years. If you mean the industry as a whole, I simply disagree. The talent pipeline is not there anymore. You used to be able to get started without an Engineering degree by “getting in at the bottom.” Now somebody on this path will never get past power supply swaps because they’re just doing drudge…

We is engineers who work in datacenters. Colocation datacenters have existed for several decades and remote hands working there would probably not see much difference between getting a call/fax/email/app telling them to do a break fix order. Having worked in dozens of datacenters over the past two decades is that there will always be newer semi-technical workers there doing break fix. A lot of those people would consider those jobs dead ends too and so this cloud datacenter thing is nothing new, its just on a much larger scale. The datacenter operators don't have to keep working there to level up their careers but working in a datacenter gives familiarity and the confidence to try more technical roles later.

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

#148

Post by bluedino »

Let's look at it storage wise. A 480TB Backblaze 6.0 pod costs $22,000 to build. Dropbox has 90% of their data stored on their custom systems, and they have 'multi-exabytes' of data.

Let's just say they have 5 exabytes. A petabyte costs roughly $50,000 using this scenario, and 5,000 petabytes would then cost $250 million dollars alone.

If they had 20 exabytes...they would have a billion dollars in storage hardware alone?

This doesn't include power, networking, labor, none of that.

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

#149

Post by chris11 »

evrydayhustling wrote: post #105
dswalter wrote: post #81

Earlier quoted context omitted.

Why must Amazon pass off its savings to your org, in this example, rather than taking most of those realized R&D savings to their own investors? Sure, AWS may be constantly innovating on doing things more cheaply and scalably. Sometimes, those savings get passed off to the customer; but not most of it, and not necessarily all that frequently.

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…

I was at an F500 that spent a lot of money with AWS. I heard rumors that services were becoming more expensive. If you are spending a ton of money, moving off of AWS is major project that will require a lot time, money, and personal. You are kind of locked into AWS unless you convince senior management you should move off.

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

#150

Post by filmgirlcw »

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…

> My biggest question is whether cloud providers could achieve a scale where they are able to offer the most optimal infrastructure costs for specific businesses. Maybe this is the case for smaller or mid-size companies, but I'd be interested to see where the inflection point lies.

Disclosure: I work at Microsoft on Azure, but I’m on the product/dev tool side not on infra.

I think this is already happening to a certain extent and will happen more in more verticals as time goes on. There are massive government use cases for the cloud and it isn’t as if governments and agencies haven’t been maintaining their own datacenters and servers before. Clouds optimized for healthcare are also a thing and are only becoming bigger — again, industries that have long maintained their own infra.

You also have the private cloud model, which OpenStack pioneered but Azure Stack and AWS Outpost have put their own spin on, which essentially lets you host specific cloud services and tools on your own infrastructure.

There are always going to be some businesses that reach a size and scale where it doesn’t make sense to offload to the cloud, where paying for people to do maintenance and support, build out monitoring, handle everything soup to nuts makes sense. I think Dropbox, which is a storage provider, is a key example of that.

I talked with the then CTO of Dropbox right after it finished moving from AWS to it’s own datacenters and the process was extraordinary and really impressive. For what Dropbox is doing, it makes sense that it owns and operates its own infrastructure and storage and tooling.

Of course, you can also have the inverse. Zynga famously moved off AWS as its demand peaked and it saw the cost savings, and then had to move back to it, after demand died down and the numbers of owning and maintaining its own infrastructure no longer made sense.

Netflix has moved much of its stuff in-house, but still relies on AWS and likely will for quite some time.

But on the whole, yes, I absolutely see cloud providers moving to offer specific business and business vertical centric solutions with pricing that is lower than what those businesses could achieve on their own, even if you take some of the “services” stuff out of it snd are just looking at raw infrastructure costs.

Post reply on HN