Hacker News Reader

Comment by dcchambers | Hacker News Reader

I have some personal stuff hosted on DO. I really like their options, service, their branding, UX/UI, etc...but they are kind of in a weird spot. Halfway between being good for cheap personal projets, and being good for enterprise.

If I want a simple VPS there are cheaper options.

If I am an enterprise spending millions/year on cloud infra I am probably only looking at AWS, Azure, GCP, etc.

How does DO get out of this spot? I want them to succeed and I will continue to support them as the big guys need the competition, but I fail to see how they compete against the likes of Amazon without undercutting significantly...and that won't bring profits. I think the margins on cloud infra is already pretty thin.

View on HN

You hit the nail on the head, we are best for SMBs and teams that want to get things done quickly and don't need the hyperscale and added complexity of AWS. Our focus has always been on simplicity and as our customer needs and our own internal needs have grown we've added additional products to continue to allow customers to scale with us. We launched with just Droplets in 2012 and have since added block storage, Spaces object storage, load balancing, kubernetes, managed databases, firewalls, and more.

Our goal isn't to be bigger than AWS or Google, but simply to provide a great service to our customers and to continue to expand our offering as those needs grow.

Simplicity and reliability are the biggest reasons I've stayed with DO for so many years. The combination of droplets, spaces and simple load balancing is the sweet-spot for me. Not to mention a simple, consistent bill each month that makes budgeting much easier. Thanks DO and keep up the amazing work!

yeah I'm not sure about your last paragraph. when I interviewed at DO, the few business folks I spoke to were adamant on taking on AWS. I kind of scoffed but hey I'm not running DO.

"Business folks" :)

But all kidding aside, the numbers just don't make sense. AWS is doing $30B in revenue and has tens of thousands of engineers.

Google which many can argue has some of the best engineering on the planet is going after AWS investing billions in datacenters and again using thousands of engineers.

So to think that DigitalOcean with a $300MM debt line and 250+ people in engineering is going to go after AWS just doesn't add up.

Certainly when you have 500+ people in a company everyone has different opinions and view points on where a company should be headed, or who it is competing with and so forth, but for me the priority has always been clear. Focus on the SMBs, and developer teams, that need the flexibility that AWS offers without the complexity because you aren't managing a $10MM hosting budget.

I don’t necessarily disagree with you, not sure I’d place a bet on DO taking on AWS realistically, but:

> So to think that DigitalOcean with a $300MM debt line and 250+ people in engineering is going to go after AWS just doesn't add up.

Lots of Money + Lots of People =/= Guaranteed Success

It’s more a matter of taking on the market from a slightly different angle.

One viable path I see is for DO to ramp up their enterprise appeal by offering services on par with AWS, while cutting down the operational complexity of managing AWS services (using AWS = incredibly high overhead with configuration and everything else).

> Lots of Money + Lots of People =/= Guaranteed Success

I'll agree on that.

I won't agree with the rest though, I don't think DO can compete in an enterprise level.

Actually in the longrun I don't see DO ever being viable. AWS and google are going to become simpler as the time passes and they are going to take over that margin that is left on services like DO, ovh etc. I think the only viable business plan for them right now would be to focus on their simplicity and continue building on it hoping they will get bought out by the big players. Obviously thats not good for us -> amazon and google owning everything on the web but unless there is some crazy law that limits them from doing so, they are going to do it.

There are still MANY old style ISPs playing in the DO-like space. Hosting is HUGE. Plenty of room for many DO like players at .1-1% marketshare.

Google and AWS are NOT going to become simpler, to the contrary, that would defeat their entire proposition. To eliminate the traditional enterprise data center entirely, build sufficient lock-in moats, and then continue to innovate around costs and value requires tons of bespoke capabilities and complexities.

> AWS and google are going to become simpler as the time passes

Really? I see them becoming more complex as they add more services.

I agree I don't see AWS and Google becoming simpler, but instead becoming more complex. They are also catering to workloads that are completely different and they really need to have every single possible knob and dial and configuration possible to satisfy these enterprise customers.

Doing what they do is complex and takes a certain skill set, but also doing simplicity right is it's own skill set.

I would say that I would be more worried in the past when AWS was more of a single player, but now that Google and Azure are firmly in the picture, AWS has some real competition in their core market which takes their focus away from us which is great.

> One viable path I see is for DO to ramp up their enterprise appeal by offering services on par with AWS, while cutting down the operational complexity of managing AWS services (using AWS = incredibly high overhead with configuration and everything else).

Absolutely. Getting into AWS, etc can be overwhelming. Getting into DO is incredibly easy.

ye I agree, it feels impossible to penetrate that market unless you have many billions and the actual engineering knowledge behind it.

I think DO is good for the small/average player that doesn't wanna invest time into having his team learning the AWS quirks because his business isn't in heavy need of it.

If AWS or even google ever goes after the small/average user by creating an easier to understand/navigate service for the user that doesn't need all that overhead and time investment of that bizarre learning curve just to setup a simple application, then its just gonna kill DO and other services like that out there.

Two words: "egress cost".

Currently both AWS and GCE cost is multiple times more per gigabyte served, compared to DO. E.g. AWS S3 is $0.09 per gigabyte served, while Do Spaces is about $0.01.

For certain businesses, traffic is a major cost, and paying many tomes as much for it could make them unprofitable.

So no, a cloud provider of the DO class must be long-term viable, just in a different niche.

>If AWS or even google ever goes after the small/average user by creating an easier to understand/navigate service for the user that doesn't need all that overhead and time investment of that bizarre learning curve just to setup a simple application, then its just gonna kill DO and other services like that out there.

Lightsail doesn't seem to have killed off DO: https://aws.amazon.com/lightsail/

I disagree, you can't just tack on simplicity to an incredibly complex product. Elastic Beanstalk is not in the same class as Heroku (despite both running on AWS). When you look at AWS the strength is in breadth and loose coupling. This allows internal teams to push a massive set of products forward in parallel, but the tradeoff is that the seams show everywhere, which directly works against having a simple and cohesive product experience. Theoretically it is possible, but I think you'd need to brand it outside AWS and probably dedicate way more headcount than a beancounter would deem reasonable to compete with DO, and even then I'm not sure AWS or Google are structurally capable of producing such a thing.

What I really wish for is a simple way of running docker containers (like AWS Fargate, or at least ECS), because I want to run docker containers across multiple droplets, but I don't want the full complexity of Kubernetes. Also something akin to auto-scaling groups. If DO had those, I'd use it a whole lot more than I do (currently I only spend use approx. $140/month on DO).

I just led a migration for my small team from Zeit Now to Render (https://render.com/). It has filled this need pretty well. There are some features that I wish existed but overall the simplicity has been great for our use case. They do not have auto-scaling but it's planned (https://feedback.render.com/features/p/autoscaling).

(Render founder) Glad to hear it, and I hope you've posted your feature requests on https://feedback.render.com. We're investing heavily in growing the team and putting a strong engineering foundation in place so we can keep adding new features quickly and reliably.

How is Render different from Heroku? It seems like a slightly more expensive than Heroku with slightly fewer features?

Render is more flexible than Heroku: you can host apps that rely on disks (like Elasticsearch and MySQL), private services, cron jobs, and of course free static sites. You also get automatic zero downtime deploys, health checks and small but handy features like HTTP/2 and automatic HTTP->HTTPS redirects.

It's considerably less expensive as your application scales: a webapp that needs 3GB RAM costs $50/month on Render; on Heroku you'll pay $250/month for 2.5GB RAM, and $500/month for the next tier (14GB RAM).

And you get free chat support.

Repobus (https://www.repobus.com/) yet to be launched will provide this functionality in DO or any other cloud. Simply add your nodes and that is it. It is a Heroku like platform on your VPS. Launching sometime in April.

Another thing I would love that I forgot to mention is hosted Kafka (or even hosted rabbitmq or similar), as I find them a pain to manage myself.

Have you checked out Hashicorp's Nomad?

Yeah, I quite like the look of it. I’m still scared of actually running it myself though since the master node requirements are a few servers. I guess I need to look into it again.

Their documented "minimum requirements" are quite ridiculous TBH. I mean, officially Vault requires 6 Consul servers (dedicated to Vault, mind you) to be considered ready for production. I doubt most companies using Vault with Consul follows this.

You I think you could be fine with 3 of the smallest machine types.

That’s basically what put me off trying nomad. Nit just the minimum number, but also their stated hardware requirements. From their documentation:

Nomad servers may need to be run on large machine instances. We suggest having between 4-8+ cores, 16-32 GB+ of memory, 40-80 GB+ of fast disk and significant network bandwidth

Basically the cost if the Nomad masters would be much greater than the cost of what would run my actual applications. That’s a non-starter for me.

> You I think you could be fine with 3 of the smallest machine types.

Maybe, but if that’s their official stance, it would make me very nervous to run a production system with lower spec machines.

They ( and many others ) better pray central banks keep printing money by the ton, because once the crunch comes these IT SUVs will be the first to go.

It really depends on how what your workload is though.

We have been running between 100 and 200 Jobs in Nomad, with the quantity of clients doubling then shrinking every day using 3 × t3.micro for the servers since years.

We have yet to see our Nomad usage increase enough to get rid of these machines.

Maybe there's some element of "geek macho" from their side here, but this is their recommendation for a supporting a "small" workload, where I suspect you find yourself on the very low end of that.

Like, you're not doing "big data" unless we're talking petabytes per day.

We're working on it.

What about hosted kafka or hosted message queues? Are there any plans (even tentative) in that regards? That’s the other missing piece that would make a huge difference to me.

One final question: will spaces support proper static site hosting? There was a ticket about it stating it was planned for Q3 or Q4 last year. but there was no follow up.

On a somewhat related note, what is the best way for startups to work with DO to get free cloud computing? My product is a search and analytics engine for github and I would like to offer a free version for people to use and I would like them to use DO since I find it has the best bang for the buck.

I can see people that would like to use my solution to not need the bells and whistles of aws, azure, etc.

Does DO provide startup assistance like Microsoft, which gives free cloud computing? My goal would be to say something like try DO free for a week and if you like it, continue using them.

DO used to have referral credits. A $5 of referral credit could feed a $20 instance for a week, of the smallest instance for a month.

Check out our Hatch program - that's probably what you are looking for:

https://www.digitalocean.com/hatch/

Where's your serverless product? I don't want to manage Kubernetes any more than I want to manage instances.

I agree. I’d like to use DO but I’m full in on serverless and GCP keeps me on their platform with Cloud Functions and now Cloud Run.

I do love the simplicity, but I think the big thing that is missing is private networking. I'd love to be able to build my own network that connects droplets together and assign them IPs, and that would let me setup VPNs between datacenter.

Also, adding at least another US datacenter would be great.

@raiyu Any more insight about this debt-raise round? Why didn’t DO raise from a VC and increase their valuation to the next level?

I've always believed that as cloud stacks mature and get commoditised various VPS providers will provide a decent alternative to being locked into the AWS or Azure ecosystem. I think the main drivers for people actually migrating will be cost and flexibility, where flexibility will be provided by software stacks that will be deployed on top of simple compute instances.

I'm a customer on DO as a small business (mobile game). I wanted some linux servers that I have control over, and I wanted a managed database where someone does backups for me. I also wanted simplicity and a nice UI. I didn't even compare prices, I just thought it was a good deal and didn't look back. At first I tried Heroku but it was too much adapting our application to their way of doing things. DO is just hyper-simple for me. So, thanks DO :)

Heroku was really built for Web Apps (especially Rails)

Aren't they just a middleman for AWS anyway?

AWS is infrastructure as a service (IaaS). You need to worry about load balancers, backup scripts, firewalls etc. Ie you need a plausible sysadmin to run it. Even Elastic Beanstalk requires some sysadmin.

Heroku is platform as a service (PaaS). It abstracts away the complexity of AWS. The base cost is more expensive than raw AWS but you don't need to source and pay for a sysadmin.

AWS, Heroku and DO aren't "better" than each other, they serve different use cases.

I have been dreaming of running my own heroku like interface on a DO droplet. Just for small sites, side projects and such.

With heroku you need to pay to keep things running, though sadly with DO you need to do a lot of sysadmin work to keep things running.

Have you tried Dokku?

After failing with Heroku we migrated to using Dokku on DO. And yes it is more work to do it ourselves, I didn't want to do that but here we are :) It was not a lot of work, dokku is really nice, but it was/is some work. I was hoping for almost no sysadmin work but that didn't happen.

Caprover's probably the closest I've experienced, but there was still a bit to do.

What... no. Install Caprover or Flynn or any of a dozen systems that provide Heroku-like functionality.

In my case I found it beneficial to just move everything to Hetzner. It offered everything I need at a lower price, and I foresee that for some of my projects I'll need to get 'proper' servers instead of the lowest-tier VPS offerings. It's easier to do that when it's all in the same ecosystem.

That said, it's only been barely worth the hassle. The difference in cost for the lowest tier, at least, is almost a rounding error in my finances.

Furthermore, it feels strange that I still often end up at DO pages for various tutorials or articles to help me set things up.

I guess mostly I still find myself fond of DO, have been helped immensely by their free articles and tutorials, and am immensely impressed that they haven't gone to shit like all the shared hosting providers I've churned through in the years before I moved to DO. and while it was ultimately worth it for me to move all my projects to Hetzner, I wouldn't particularly recommend that approach to anyone unless <10$ a month is something they are concerned about.

I guess there is a niche, not every enterprise spends millions a year. We're a pretty happy customer that spends maybe a few thousands every month, if there are enough of us it could already be a sustainable business?

Aren't enterprises that spend millions a year the niche and not the other way around ?

DO's branding and a lot of their offering is pretty good, but their locations for non-US customers are much worse than many of their competitors. For instance (and a particular point for me), they still don't have any presence in Australia after over half a decade of it being marked as "under review" on their customer feedback page.

Having geographical locations to back up the quality of the offering is a step forward IMO.

Most likely Telstra (40% market share telco) charge too much for peering network traffic in Australia? Same as how Google Cloud free tier excludes only China and Australia.

Vultr has a POP in Sydney and seem to manage just fine, offering plans similar to DO. I like DO but as I'm based in New Zealand, it's a no brainer to go with Vultr due to their Sydney POP.

I thought their $5/mo machines are the cheapest on the internet for the specs available. Are there cheaper options? I’m hosting a low traffic page that gets maybe 500-1000 views a month and even at $5/mo it seems overkill IMO.

Check scaleway and hetzner

https://www.hetzner.com/cloud https://www.scaleway.com/en/pricing/

Pretty sure Linode is comparable price wise and in Europe there are cheaper options like OVH. Digital Ocean pricing becomes pretty comparable with GCP and others when you start specing up to a "production-grade" server i.e. the kind of server you actually want to run Postgres on.

Their K8s offering I think is the cheapest of all major providers but you lose out of secondary benefits like GKE's fantastic log analysis tools (I think its called stackdriver or something).

Their database offerings are in about the same range as other providers (not comparing it to Google Cloud $panner).

Just popping in to say that OVH is the cheapest option for a reason -- support/monitoring/any other features you might want are just not there like they might be on other providers.

I host a single tiny website on OVH and they restart it randomly every ~4 months without warning (seriously). It was annoying at first until I set our services to run on system boot.

Some more discussion at discourse.org that's relevant: https://meta.discourse.org/t/migrate-from-digital-ocean-to-o...

> until I set our services to run on system boot.

Which pretty much is how things were always done, even in the PHP-CGI days.

I suppose it depends on what OVH offering you buy. I have multiple servers at OVH with multiple years of uptime, but those are dedicated ones and not the cheapest Kimsufi offerings.

I have multiple projects on the cheapest Hetzner.cloud offering and haven't run into any of these issues. might be worth looking into.

You can get OpenVZ or even KVM boxes for half the price or better. I have one running as a simple nginx webserver and OpenVPN server that I pay under $30 a year for

AWS Lightsail has a $3.50/month option that might work for you if you want a VM. You could even consider something like nearlyfreespeech.net if you dont need full control of the tech stack.

Lightsail viciously throttles to something like 5% of original CPU after only a bit of use. Not suitable for much.

For static site hosting there are a ton of "good enough" free options these days like GitHub Pages, Netlify, etc.

For pure VPS, there are cheaper options, especially if you don't need a ton of customer support like DO offers.

For example, buyvm.net has a VPS with 1 vCPU, 1GB RAM, 20GB SSD, and unmetered bandwidth for $3.50/mo. DO's cheapest VPS is 1 vCPU, 1GB RAM, 25GB SSD, and 1TB bandwidth for $5.00/mo.

Digital Ocean does offer a ton of value in other ways - support, uptime SLA, and other managed products...

Edit: Check out https://lowendbox.com/ to find cheap VPS providers.

Very happy BuyVM customer. The unmetered bandwidth really is unmetered, and support from the founder has been remarkably transparent and often minutes when he's awake and working.

Have you checked the performance of their VPS:es? Any numbers to share? I've been thinking of using them because of their anycast support.

CPU:

    model name : Intel(R) Xeon(R) CPU E3-1270 v3 @ 3.50GHz

So yes, it's a 7 year old quad core, with maximum of 32 GB of RAM. You only get access to one core (technically thread); 512NB to 2GB nodes can burst to use the full thread, but are expected to not 100% it. The 4GB node (1/8th of the server) is allowed to fully peg their thread.

Some of the newer servers are powered by AMD Ryzens, which is a great thing (they are far better perf/$ now; they're on GCP, and Tencent is deploying tens of thousands of ryzens in their DC).

I have a 2GB node, and I've ran Geekbench 5 and got a score of 661. Here are the results: https://browser.geekbench.com/v5/cpu/1278946

For comparison, a Vultr 1GB ($5/m) has a geekbench score of 2413.

For most web server needs, BuyVM should suffice.

Yes, other providers are much faster; but other providers don't offer unmetered bandwidth. BuyVM is great for bandwidth heavy, compute-low loads.

For 500-1000 pageviews per month (maybe even per day?) I think you could get away with the free tiers of (if static: github pages, netlify), or (if dockerable: Google cloud run), or (if postgres required heroku) or now.sh or lamba. vultr offers a $2.50 VPS.

>Are there cheaper options?

Definitely. There is an entire ecosystem of small VPS providers below DO & co.

e.g. I'm currently paying 7 USD for a 4 core w/ 16gig ram. Modern hardware too - Ryzen & NVme.

Trade-off is you need to spend time hunting for once off deals & there is no guarantee the provider will still be around tomorrow & you never know whether it's oversold.

>How does DO get out of this spot?

What if they package the available open source versions of AWS and GCP services and allow people to move off amazon and google? Have DO as one region and at first sell it as a backup. Then, make it enticing for people to scale it up so that they shift some work over.

DO could become the place where people can run infrastructure that mixes AWS and GCP. After some acquisitions, companies will have services in both worlds. Why not unite them in one place?

> If I am an enterprise spending millions/year on cloud infra

They do more than just VPNs. They do object storage, containers and K8. They have command line an API based tooling.

A small enterprise could do a lot with that.

If you were talking a $100 million dollar budget and you needed bells and whistles like IAM Lambda or Glacier I could agree.

Reply on news.ycombinator.com