Comment by Townley | Hacker News Reader

Comment by Townley | Hacker News Reader

I'm rooting for them as one of the best potential guardians against the cloud provider market becoming even more of an oligopoly.

Cloud resources should be a commodity. Providers should offer compute resources, persistent storage, load balancers, and MAYBE a small handful of other services.

The way Digital Ocean succeeds against AWS is by aligning itself with this idea, and competing on specialization. Forget competing with lambda; let me run my own serverless application. Don't worry about IAM; let me configure LDAP. Don't waste developer hours on service-ifying the latest NoSQL storage trend; write high-quality tutorials explaining how users can do it themselves.

And most importantly, continue to invest into open source and community resources. There are developers willing to fight the good fight against proprietary walled gardens like AWS/GCP/Azure, but it has to get easier. Configuring HA postgres is harder than paying for RDS. Paying for GKE is more feature complete than using rancher or kubeadm to make my own kubernetes cluster. This friction is an existential threat when Azure can make my problems go away for cash.

I don't know if Digital Ocean can succeed against the big cloud providers, but if they do it won't be because they made a better platform; it'll be by playing a totally different game.

I'm rooting for them too, for the same reason. It's unhealthy for everyone to have Amazon be the only real option for servers. Just look at what happens when Amazon goes down for an hour. It's not hyperbolic to say the entire world notices.

My work doesn't even use Amazon, but when they went down last time, every _other_ service I was using used Amazon, so it didn't matter if I was vendored in or not, I had to just leave work and call it a day.

DigitalOcean is a solid platform and I use them a lot. I really hope they are not only able to succeed but bite into the profits of Amazon and the like.

Unless DigitalOcean creates their own equivalent of a GovCloud region, they won't be an option in my industry. I do still like using DigitalOcean and Vultr for personal projects and the like.

> It's unhealthy for everyone to have Amazon be the only real option for servers

I'm not belittling AWS' commanding market share, but they are hardly the "only real option for servers." Google and Azure are both in the leviathan league and have competitive pricing.

Google has the worst customer service I've ever heard of and people get wrongly banned from Google Cloud with no recourse all the time. It's not a real option to do anything vaguely important on.

Not my experience at all. We've been running in production on GCP for almost four years, four k8s clusters and a couple of hundred vms handling ~20k rps at peak, and their support has been excellent. Most importantly we haven't needed it very often.

Not my experience, either. Not a large account by any means, and support is excellent, both from a technical and business point of view.

> people get wrongly banned from Google Cloud with no recourse all the time

[citation needed]

There have been a few isolated cases that generated a lot of bad PR. From what I've heard, there were significant changes in how abuse is handled to address the root cause.

AWS has their own competitive service against DigitalOcean called Lightsail. Their cheapest plan is $3.50 month (512MB/20GB SSD), the bigger plans are roughly similar in pricing and features to DigitalOcean.

The irony of this article is that current customers like myself who read this news are now looking at alternatives.

I spend roughly $120/m with DO and have done for the last 2 years. I have a majority 5 usd droplets and 1 20 usd droplet.

LightSail may be the way to go for now.

Here comes all the migration work.. Fun :)

Last time I touched Lightsail it looked like a t2 instance rebrand without the option to pay for extra CPU credits.

It grinds to a half if you use up your compute stipend and you cannot do anything about it. That's a big risk for a VPS-like use-case.

t2.nano is fine for a low traffic website, which is what most Lightsail instances will be.

Why migrate? Afaict, Digital Ocean isn't going anywhere, it's mostly internal restructuring. I'm sure you posted before the comment by one of the cofounders, so definitely give it a read before you make any decisions.

Personally, I have more instances on vultr and linode, but I still keep an instance at DO so I have a reason to keep tabs on them. I have no qualms about recommending their service to others.

Yes, I did indeed post this before the comment from the co-founder.

Sticking to DO!

Do yourself a favour and compare the performance of DO and Lightsail before you migrate. It may be the same or similar in pricing, but every benchmark I've seen shows worse performance on Lightsail. On a value for money basis, Lightsail is most likely more expensive.

>...against the cloud provider market becoming even more of an oligopoly.

I'd be interested in the more opaque aspects of all cloud hosting i.e. Unfettered, and or, at least, Invisible government(s) access -- however there is a really dark double-edge on that privacy hope. (specifically, that there is a lot of nefarious dealings on dark web systems which humanity would be better without)

Its almost as though we also need the antithesis to the dark web. Whereby, if we were to consider the contemporary Internet as the 'Gray Web', the ostensibly-perceived-as-criminal 'Dark Web' the host of Nefarious Dealings, and a 'White Web' for things on the transparent, or at-least validated on the up-and-up.

I would propose that ALL sites with content directed at children must be regulated on the 'White Web' (looking at you, YouTube >:-[

Anyway - my overall point is that Cloud has become 'Privacy-out-of-sight-out-of-mind Land'...

> I'm rooting for them as one of the best potential guardians against the cloud provider market becoming even more of an oligopoly.

So much this

At $275m in sales they're already a drop in the AWS/GCE/Azure buckets though...

Sure, but they're still big enough to maneuver well against competitors. I see them competing more with the likes of Vultr and Linode than AWS/Azure/GCE.

I think they would benefit a lot from adding a thin Heroku-like PaaS layer while keeping the option to "build from scratch".

That’s why they acquired Nanobox, presumably.

Never heard about Nanobox up until now. Interesting, let's see what comes out of that.

Dokku (and other things like flynn) get you that heroku experience, but you do a bit yourself. Following the same idea as what the parent said

They're working on this. I was just asked to do a survey about my "team's" PaaS usage and offered an invitation to a working group scheduled later this month.

> let me configure LDAP

I just got a nervous tick. Having done it before IAM is the killer feature of AWS.

I'm rooting for them too. I have a bunch of servers running various workloads. They have been great and I love their product. I really hope they are around for the long term

Reply on news.ycombinator.com