Ask HN: What Happened to Digital Ocean?

>Not too long ago DO was touted as the #2 cloud behind Amazon [1]

Is there a time of where that was said in the podcast. I could bet money at no point in time was DO no.2 in Hosting or Cloud Hosting in revenue, profits or customers. They might have been the case if you count instances / droplet or Domain name, but both of these are quite irrelevant.

They have growing pains, as with ALL companies does, and they reach a point where growth has slow down due to market interest and competition. Not only has Linode becomes better and more competitive thanks to DO, there are also new comers into what I considered as Middle Class Cloud Hosting such as UpCloud and Vultr. You also have Giants ( in their own right ) such as OVH and Hetzner.

>Its not like the market for their stuff is that small, Azure and even GCP have proved that.

No one gets fired for buying IBM. That mentality is the same for Intel ( vs AMD ), using SalesForce rather than other CRM or ERP, and across All industries, from toilet paper roll, food, water, cleaning solution etc. Big Enterprise wants to works with other Big Enterprise. If you have ever sit in the chair of the purchasing, or Supply Chain director you will understand the risk of working with smaller companies is far too great. DO aims at people who are willing to take that risk, SME or startups. And that market is fundamentally different to Azure and GCP serves.

So your answer aren't really technical, but more on the economical and business side of things.

yes, there was those exact words were said. at 15:30, peter levine (a16z GP) says 'recent studies show you guys are the second largest cloud provider by number of server instances, second only to Amazon'

I'm still pretty new to cloud - do you mind explaining why number of instances is not significant? sounds pretty significant, it indicates active, paying usage i think?

> So your answer aren't really technical, but more on the economical and business side of things.

i can accept this. but are all "new cloud startups" doomed this way? (I work at Netlify so I have a vested interest in this haha). At some point there's some sort of Glass Ceiling where you run out of everyone willing to Not Buy IBM.. and then the Big 3 clouds just clone you and run away with it? seems pretty bleak for the risk/reward of innovation/VC money.

>'recent studies show you guys are the second largest cloud provider by number of server instances, second only to Amazon'

It is just like any theory of economics. You can have a huge number of $5 Droplet ( aka Instances ), but limited number of high end. Imagine of you have a Million instances and 90% of those are $5, that is only $4.5M monthly revenue.

Having said that, I seriously doubt that studies to be correct. Because there is no way of knowing AWS's true number of instances. And that sentence actually said "serve out", which means they could be counting instances or Address that serve directly to the internet. i.e Excluding instances that are not served on Internet, or background instances. You can spin whatever way you want, by the only single source of truth ( At least at this point in my life and current knowledge tells me, ) Revenue dont lie. You can muddle around with Margin, Net Income etc. But Raw Revenue speaks volume.

Another point worth mentioning is these sentences are often without context, much like the recent headline "Dow has the largest drop in history", ( or a slightly better headline would be) "Dow has the largest point drop in history". It is irrelevant without knowing the percentage. Dropping 10,000 point when Dow is 1 million would only equate to 1%. Not to mention A16z is a board member of DO, so a clear conflict of interest and remember to read it with cautious, as they are designed to trick its reader into thinking, for this instances ( no pun intended ), DO is the 2nd Cloud Provider. So they could market themselves as big enough for many other customer.

> i can accept this. but are all "new cloud startups" doomed this way? (I work at Netlify so I have a vested interest in this haha). At some point there's some sort of Glass Ceiling where you run out of everyone willing to Not Buy IBM.. and then the Big 3 clouds just clone you and run away with it? seems pretty bleak for the risk/reward of innovation/VC money.

Absolutely not. History has been clear most big companies fail to innovate. I will point you to a tweet from the same A16z BE [1], "Zoom, Shopify and Stripe are three fun examples of companies that were obviously impossible because the 'tech giants can easily expand into new areas and squash competition'. And Zoom wasn't even doing anything 'new'. Combined value of $123bn."

And for DO, they are still doing great. They are growing, just not as fast as they were, they are not aiming to be AWS or Azure, they are perfectly happy within the 2nd Tier in Cloud Hosting. They are expanding SaaS model with Managed Database. The market itself is still growing, there are plenty of space for everybody, including Netifly. Which in its own niche is possibly the best there is.

No one has even reached or seen the ceiling yet. Not for another 5+ years at least. And when that point is reach, you pivot into other relevant thing. Apple didn't start by making a Phone.

[1] https://twitter.com/benedictevans/status/1234622648363163649...

great answers, thanks very much!

Maybe the cloud has bifurcated and DO is #1 within the "light cloud" market which is a lot smaller than the "real cloud" market.

thats an interesting concept. care to develop it further? how to define the upper limit on what is "light cloud"? is it defined by traffic? bandwidth? horizontal scaling? team size? ARR?

I was thinking of advanced features like IAM, CloudTrail, direct connect, etc. that are only used by the most sophisticated customers.

That seems accurate, and that's actually a pretty big market.

For me, DO has been great. The apps I host don't have a lot of heavy traffic and run great on a $20 a month VPS. I have 3 of those running to spread the load. One serves the app files, one holds the database, and another handles email and DNS.

For the most part, everything I've done there has worked as expected and I've not had any serious issues with downtime due to DO internal problems.

I've been using hosting services for over 20 years now and DO has been leaps better than any of those I used previously. That's not to say Linode or others aren't great, or even better than DO. Since I've not used any of those mentioned by others here I can't speak to that, but I do wonder if those who dump on DO may have pushed their setup a bit too far.

Many of the complaints I see are about having issues with WP. The few times I looked at that I came away thinking it was an overweight convoluted mess. One might be able to squeeze it into a $20 per month VPS but that doesn't mean it's a good choice on the DO platform. And on my own server logs there is a constant stream of requests looking for it and I have to suppose they're seeking potential exploits.

So, for me, "light cloud" is exactly what I need and DO does a great job of providing it as a service.

great answer. thanks for sharing!

This is an interesting one for me. I run a small network of sites, around 130 or so (Wordpress) that used to be hosted on DO. We had CDN and heavy caching implemented, as well as fail2ban and other security features. For some reason, every site would go down at some point monthly: error connecting to database, Apache crashed, etc... couldn’t get any support on it.

We switched to WP-engine and haven’t had a single issue. We now host our web apps with Azure without issue as well. We spend maybe $5-$7k / month on hosting (which is a drop in the bucket for larger companies), but I can’t imagine we are the only ones who encountered this.

that sounds awful. i would certainly be interested in hearing from someone at DO about the real behind the scenes on why it was so bad.

It sounds like you would have stuck with DO if they were simply more reliable though - so their losing you is entirely their fault.

I had the same story, my WP site would go down everyday. Switched to SiteGround and haven't looked back

Among the lesser clouds, I think Linode is better than DO if you just want a reasonably priced vps. I ran dev and CI workloads there while running the prod stuff on AWS and was always happy with them.

We stopped using it when we were always the first to notify them when they were having data center issues...always a switch or gateway or some other excuse causing major network issues.

Not to mention their security model allowed any instance to talk to another instance from any account until 2019....

would you say you outgrew them? bc is seemed like their value prop has been fine (for small usecases) and consistent since the beginning.

or would you instead say they mis-executed - if they were only better on their network issues and security model you'd have stuck with them all the way?

It was a bit of both. I don't think they had the infrastructure to support everyone, I think they had major growing pains as they tried to scale, and I don't think they had the experience to identify sources of partial-failure proactively which resulted in 40-360 minute effective outages that weren't as easy to fix as with AWS where we can shift the stack to a new AZ in minutes.

DO had no cross-"AZ" networking. We tried to form a tunnel but it was so unreliable they admitted they needed their own backbone to support that use-case but even when that did come out we'd still have to use public IPs...no VPC peer type thing. Our only option to switch "AZ"s was to have downtime and that wasn't acceptable.

We were able to increase capacity 300-500% on AWS for the same cost so there's a lot going on in the equation for us. We needed better capability to dynamically resize and more options for low-cpu high-memory. We needed better flexibility with disks especially by network like EBS but given their network issues I understand why they don't have that.

this is all really good feedback/food for thought for me as I learn about what matters to people. thanks!

personal anecdote : I quit DO for Linode after a string of fairly hostile customer service interactions.

After having switched, I realized that Linode provided a better value for what I needed.

As with most poor customer service interactions, I doubt I was the only person who experienced such things from them.

Curious what was the issue? Was it more technical or billing related?

Long time Linode user, moved to DO last year. I can confirm that the quality of support at Linode is really good- you interact with people that care and know their stuff, follow up and through. My interactions with DO support left me with the impression that it was a generic outsourced service were “complaints” go to die or get killed with indifference

Same. On multiple occasions. I will never risk hosting on their platform to save a few dollars. No issues with any other cloud provider. Been using linode, AWS, GCE for years.

Ye, right, the DDoS and security incidents Linode had are absolutely no issue.

> How do you go from one of the most hyped IaaS companies to suddenly losing steam? If it worked for them from 10m - 100m ARR, how do they lose steam from 200-275m? Its not like the market for their stuff is that small, Azure and even GCP have proved that. I'd love for some perspectives from people who know this market, because it's baffling to me.

DO's market (raw compute and occasionally unreliable object storage) and that for Azure, GCP, and AWS (top tier cloud ecosystems) are entirely distinct. The former competes on price and is a commodity, the latter are premium offerings that large companies happily write checks for.

thanks for putting it like this, i guess I could see that.

I'm also interested in the "new Heroku" startups (Render, KintoHub) that I guess are different enough from DO that it's kind of offtopic to this conversation. I just care as someone just starting a career in this industry.

At one point I was using DO for everything but then I got to the point where spending time on server config just wasn't productive so I moved everything over to AWS Elastic Beanstalk and RDS and never looked back.

I know DO do now have dedicated DBS products but really feel they're missing a trick with nothing similar to Elastic Beanstalk or GCP App engine - this is exactly the kind of products smaller dev teams and one-person operations need, and these are most likely the kind of people who would be willing to give DO a go over AWS and its perceived complexity.

IMO what DO offers is ahead of Elastic Beanstalk and GCP App engine in many ways. Instead of offering a single service they have opted to provide easy to install images for deploying services. These days if I wanted to set up a small service on DO I'd simply go to the DO marketplace and click a button to launch a docker instance. For a larger or more complex service I'd probably opt to use DOs kubernetes support, it's similarly a couple button presses to boot up a cluster.

In either case I can then just point kubernetes to the right address and get going with spinning up my application.

It's a bit less polished than what elastic beanstalk or app engine provides, but it is a lot more flexible without being more complex. If anything it's a lot simpler than any of AWSs offerings because DO actually has a sane dashboard unlike AWS.

I can't comment on app engine since I haven't used it but I can heartily say that the last day I need to touch that abomination of an AWS dashboard is the day I'm having a party.

I'm curious about the following two aspects: 1) were changes in DigitalOcean's executive management (in particular, their CEO replacement) prompted by founders' desire to change their focus / reduce their workload / whatever or pushed by some of DO's investors; 2) why DigitalOcean is an LLC (to the best of my knowledge, an absolute majority of VCs insist on their potential portfolio companies being a C Corporation (mostly, a Delaware C-Corp). Thoughts?

DigitalOcean was actually a C Corp, they switched to a LLC down the road.

Why would they do that? What was their investors' reaction?

I have honestly no clue, seemed weird at the time to me as well.

There's also the "Digital Ocean Killed Our Company" event last year https://news.ycombinator.com/item?id=20064169 that seems to be fresh in people's minds, but honestly that seems like a bad luck one off rather than something endemic in DO.

Sorry, DO was never considered in the same class as "true" cloud providers like AWS, GCP, or Azure. DO is primarily a VPS provider with some other services. It is in the same class as Linode and perhaps Vultr.

Abuse. I see so many bad interactions from their network that I have blocked them entirely.

Bots attempting credit card fraud are common.