Vultr vs Digital Ocean

Nearest Digital Ocean datacenter to New Zealand is Singapore.
Opened in 2014. Probably with hardware as old as Vultr Sydney.

Always try to sit as close as possible to the visitors.
And if it is possible, the customer's country (latency, law in case of webshops)
If your choice consists of the above 2.

Then I recommend Vultr Sydney.

You can try it out step by step and scale up if necessary.
You pay per hour so testing does not cost much.
If it's not fast enough then Cloudflare on top of the Vultr Sydney based server as extra.

If it does not offer satisfaction.
Then rent a server at New Zealand.

Also not satisfied?

New Zealand based server + Cloudflare

And if that does not work well.
Digital Ocean Singapore server or Digital Ocean Singapore server + Cloudflare etc.

 

Vultr Sydney does seem the one - I've had 100% uptime in the last 8 mnths and ping from NZ is 30ms. But the big "unknown" with vultr is their hardware or raid setup (they are very secretive), as you've discovered.

sitehost.co.nz seem reasonable for NZ vps (haven't used them) - they use vibe communication / rackcentral / I believe - resources / international bandwidth is low in comparison to vultr/DO - as you'd expect. I believe Ransomit also use vibe.

Some others in Sydney - latency from NZ should be no more than around 40 - 60ms:
BinaryLane; Exigent - (no experience with either). Mammoth more expensive but offers management, if that appeals.

Serversaustralia have dedicated servers in Auckland for $200mnth - but that'll blow your budget!

OVH just launched dedicated servers in Sydney too... $130AU mnth - but again, probably over your budget.

 

Assume that all of Linode, Vultr, Digitalocean and clones won't use RAID.
Simply to expensive in the VPS SSD budget market. In which they operate.
If you want to earn money as company in this sector, you need to do profit optimization and thus save costs.

They for example also don't indicate whether they use Enterprise SSD or consumer SSD,
which servers (enterprise) or home made.
You simply don't know.

I suspect they have the same business model as OVH and Online.net.
And that is create your own server with both enterprise (CPU) and consumer (SDD, non ECC memory, non RAID) components and then place the whole floor full with thousands and thousands servers.
(Scale size advantage).

Then hope for customers who do not consume much and so oversell the VPS servers a little bit on each node.

I am one of the first hour users from the above companies.
And have often lost data after hardware issues.

Which obviously is not a problem for me, as i am using backups.
But as it should be easy to fix for the company if they really used RAID.

Simple situation. You get a ticket. Problem is reported to you.
Problem can not be fixed. Data lost. Migrated to spare node.
Advice from company to restore backup.

Solution you get a few credits to soften the problem and keep you as a customer (reasonably) satisfied.

 

Indeed, that certainly would seem to be the case. However, in defence of DO / Vultr etc, they do appear to at least have the resources to carry spares, hot swaps and the like..... even if they are 100% focussed on profit - lots of downtime does not equal lots of customers....

With regard to Vultr, in an earlier post you recommended them? What's your experience been? Have you actually lost data?

Ransom it do indeed seem to get a lot of good press, but is it a one man band (Oliver) - or is it a larger company with a team?

 

Switching hard disk via hot swap or repairing the current hardware node with spare parts, is to time consuming. They almost always migrate you to a spare hardware node and afterwards ask you to restore a snapshot. The node itself will mostly be replaced on a quiet moment. But as a customer you do not notice that because you have been transferred for a long time.

They grow so fast that you mostly do not notice anything when you create a new server because you are most likely on a new hardware node.

You only notice when you are on a server since for example 2013 with an old hardware node. Their old server farm is getting old, thus fault-sensitive.

Had once 3 times crash in 3 weeks on 3 different nodes at Digitalocean AMS3.
All hardware nodes from launch = end 2013. So servers are almost 4 years old.

I prefer Linode and Upcloud.
But as the preferred location is not available at Linode and Upcloud. Then Vultr.

Jup they lost my data. But no problem. Monitoring + backups on my side.
You know what you get from a budget provider.

 

Fair comment, as you said, best protection is good DR options! The reason I ask is because right now I am looking at a new host for my current VPS set up. Current host is withdrawing from the market.

OVH new Sydney servers look attractive, but I've no experience with dedicated servers - nor with OVH so I am more inclined to stick with VPSs and instead to set up failover between hosts e.g. vultr, binarylane.

But that's complex too in terms of data syncing and another layer of complexity to maintain! I realise there are more expensive "reliable" alternative hosts, e.g. Crucial's HA Blade servers - but it is difficult to justify such price differences ($40 Vultr vs $300 Crucial). Vultr is so attractive, if I can only nail the load balancer!

It doesn't help that I have to use cpanel as that's what my clients demand. There's no easy way to mirror the cpanel set up!

@viralpoet - have you made a decision and how's it going?

 

Could be overselling.
I notice it more often at Digitalocean and Vultr then before (launch Vultr and Digitalocean around 2013)

Maybe you could do some bench-marking.
For example: Linode - Thank goodness for glances - cpu steal

 

Even Upcloud is a little, little bit slower then at launch.
But that's shared (VPS) hosting.

New nodes where almost empty at launch.
They are filling their new nodes so that's logic.