New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
I would not use ancient OVz6 technology with ancient 2.x kernel.. and it's performance just similar to HS storage, poor disk performance..
I don't know but I am not expecting robustly performance with the price I pay. Serves my needs when I pay that low and can't pay more if I am happy to get service at lower price and satisfactorily, what kind of things do you do on your storage vps?
OVZ6? OVZ7 has been out for years
A lot of usage, one of them used for NextCloud..
.
You are right, it is ancient and obsolete technology
What is ideal for one person may not be ideal to another. You wouldn't be able to fulfill everyone's craving.
Are you saying all OVZ7 VMs have bad performance? This is simply not true. Here's a benchmark from a highend VPS running OVZ7
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
Yet-Another-Bench-Script
v2021-06-05
https://github.com/masonr/yet-another-bench-script
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
Wed Jul 28 04:59:54 UTC 2021
Basic System Information:
Processor : AMD EPYC 7302P 16-Core Processor
CPU cores : 2 @ 3000.000 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 2.0 GiB
Swap : 0.0 KiB
Disk : 49.1 GiB
fio Disk Speed Tests (Mixed R/W 50/50):
iperf3 Network Speed Tests (IPv4):
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Clouvider | London, UK (10G) | 914 Mbits/sec | 926 Mbits/sec
Online.net | Paris, FR (10G) | 883 Mbits/sec | 917 Mbits/sec
WorldStream | The Netherlands (10G) | 843 Mbits/sec | 932 Mbits/sec
Biznet | Jakarta, Indonesia (1G) | 627 Mbits/sec | 328 Mbits/sec
Clouvider | NYC, NY, US (10G) | 485 Mbits/sec | 514 Mbits/sec
Velocity Online | Tallahassee, FL, US (10G) | 687 Mbits/sec | 515 Mbits/sec
Clouvider | Los Angeles, CA, US (10G) | 18.9 Mbits/sec | 570 Mbits/sec
Iveloz Telecom | Sao Paulo, BR (2G) | 346 Mbits/sec | 285 Mbits/sec
iperf3 Network Speed Tests (IPv6):
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Clouvider | London, UK (10G) | 714 Mbits/sec | 899 Mbits/sec
Online.net | Paris, FR (10G) | 789 Mbits/sec | 913 Mbits/sec
WorldStream | The Netherlands (10G) | 841 Mbits/sec | 897 Mbits/sec
Clouvider | NYC, NY, US (10G) | 299 Mbits/sec | 478 Mbits/sec
Clouvider | Los Angeles, CA, US (10G) | 163 Mbits/sec | 332 Mbits/sec
Geekbench 5 Benchmark Test:
Test | Value
|
Single Core | 1012
Multi Core | 1858
Full Test | https://browser.geekbench.com/v5/cpu/9037166
You are wrong your result are not storage VPS, I am referring T4VPS and HS storage vps performance not virtualization technology performance
I've got a VPS with 10TB storage from Hosthatch for $10/month (had to pay $240 for two years upfront, but given they've been around since 2011, this didn't seem like much of a risk). Pretty happy with it. Disk performance isn't as good as some of the other storage VPSes I've got, but for the cost I can't really complain.
I feel like I won't see any other host that offers 10TB space for that price for a long time, if ever.
1TB HDD
2GB RAM
CPU - whatever is in that offer
~500GB bandwidth will do
1 IPv4
Speed >= 400 Mbps
Preferably KVM
Non-unknown and professional provider
~20-30 Euro per year (at 2-3 yearly pricing)
.
Usually I go for specs like:
2vCPU ~700 GB5 single (EPYC Zen2/E5 v6) or 1vCPU ~1200 GB5. (E2288G/5950X etc.)
4GB Ram
50GB SSD with 10k+ IOPS
1Gbit 3TB+.
I also require that provider has floating IP option. Hourly pricing option is a very big plus.
Apart from that things like good support, high uptime etc.
it is available for me?
Things I usually look for:
Things that I do not really give a shit about:
Of course, it very much depends on what the VPS will be used for, but generally I feel that most people have totally insane demands when it comes to what they actually need.
I think you are forgot, we buy a lot of server only to idle..
You can use our IPv9 VPS to satisfy all your idling needs (out of stock until next summer, but free trial is available).
It has pre-installed YABS and an "idle" command, which is all you need for idling.
Use promo code CHOCOLATE-IDLER for double bandwidth.
1cpu
256 mb ram
15gb hdd/ssd
500 gb bw
.99 usd/m
Easy to deploy Kubernetes clusters at LET prices?
2 vCore Ryzen
3 GB Ram
60GB SSD
3$/month
Yes
I want
32 Dedicated vCores
512GB DDR4 ECC 5000Mhz Memory
10TB Full NVMe Storage
Unlimited Bandwidth @ 40gbps (No fair use policy)
/22 v4 Subnet + /32 v6 subnet
For under $1/yr
And people still claim to need 1Gb ports and nvme drives? Strange.
Why not?
I agree with this. It is just capacity. 100 Mbps and 1000 Mbps is just as fast, just that the 1000 Mbps has more capacity. My VMs usually only uses 1-2 Mbps, and having 100 or 1000 Mbps would not make a difference what so ever
It's like storage. If you only use 1 GB, what good is having 100GB or 10000GB? It's just more capacity
Yeah, I hear you. Of course, it depends a lot on the type and load of the server, but most of my VPS's rarely exceed 20Mb/s in bandwidth and everything is kept in ram, so I really don't see the need for 1Gb ports and nvme drives. I'd much rather take a 50Mb/s connection with low latency and stable storage on whatever medium is cheaper than paying for a lot of performance that I will never use.
Sometimes I do of course really need high performance servers, but then I usually go dedicated, not low end vps.
Two big problems I see
At the end of the day providing/selling VPS is the same for each and every provider; there's costs for resources (e.g. hardware, connectivity, staff) which need to be recovered plus some profit margin. If someone wants more of anything, e.g. vCores there are only two ways to make that happen, (a) really double the requested resource and ask a higher price -or- (b) make the requested resource physically smaller (e.g. % of real core per vCore) or oversell (or both).
Important side note: You want your provider to make some profit because if he doesn't products and/or service get worse plus he loses motivation.
There is no magic bullet and there is only relatively few factors that differentiate any provider from others, like e.g. location (which can have a huge impact on costs), experience (which allows for better service and support and/or using up less resources/staff), level of personal connections in the market, and financial resources.
Pro tip: the best way to save on VPS is to be realistic and to go for the specs one really needs (as opposed to number greed).
Personally I repeatedly found NexusBytes / @seriesn to be close to the sweet spot in terms of real performance and service per buck and at relatively low prices. HostSolutions for example may seem to offer the better deal but at the end of the day you get more performance - and reliability! - and better/faster support per $ spent from seriesn, @contabo_m, or @Francisco.
Sure, we can do that when we open our luna location.
The subnets won't be accessible from earth, but we sell discounted flights to the moon, so that you can go there to access your server.
Visit our sales office across the street from Starfleet Headquarters, ask for Dean K.
Chia mining allowed.
HA, useful locations and a firewall
Really want to see a very low end offers for 3-5/year with the best specs possible.
I believe @jsg pretty much nailed it on the head.
I've been around the block before. We've all seen these threads come up every year/two/whatever.
This thread is good to know what your customers want and a target pricing they have, but it shouldn't be something you focus on. Customer's end goal is to just get the best service for the lowest cost, that's going against your (as the provider) goal of least amount of resources for the highest price. The market segment you're targeting is the most price sensitive segment (this entire forum was built around 7 dollars as a threshold). But I wouldn't put it on the customers to "gauge" how much of a profit the provider should make, that's on you to decide what your time's worth.
Let's do a quick rundown on the publicly available information and look at Boomer Host as a case study on what options are available.
DISCLAIMER: Before you jump in, recognize that these are estimated numbers with information that are publicly available. My estimates could be wrong and this entire analysis can be wrong, but at least this paints a potential picture. Their internal bookkeeping and operations is their own deal. I don't mean any disrespect to anyone. I am simply a person in academia with a calculator and a napkin.
DEFINING THE SITUATION
Costs
6. 14 dollars/monthly - ClientExec unbranded (self-host) license
7. 375 dollars/monthly - Colocation (Power, Network, Space) fees (This is a higher number from DataIdeas ColoBay pricing calculator using: 1U, 1 Gbit port speed, no IPs, 1A @ 120V, no redundancy in power. You might be able to get a cheaper/more affordable deal by directly contacting sales.).
8. 100 dollars/monthly - IP Costs (Assuming renting from someone. This number is just a general number I pulled for a unit cost of 50 dollars per /24. This is a low cost estimate as some target 100 dollars/month or more per /24.)
Figure 1
Total cost comes down to:
$5,675 one-time
$448/monthly (or $5,376/yearly)
Note: This is a conservative estimate with assumption that decisions were made to reduce initial investment but instead will have you pay a lot over time (leased licenses vs owned licenses). Also, this doesn't factor in taxes, LLC incorporation fees, or anything else that drains money.
Income
Figure 2
Fund Balance
For the first year of operations, we have an estimated -$3,926.
Assuming all other costs remain static and they don't have any further growth, by the second year we should be at -$10,658. Their expenses are higher than their income by $2,679.
Updated numbers with new/updated data.
Assuming all other costs remain static and they don't have any further growth, by the second year we should be at -$2,177. Their fund balance works with a "profit" of $1,749 yearly.
STRATEGY
Well in a business there's two directions to take this:
For the purpose of this analysis, I'm assuming costs have already been minimized as much as possible. This is only a single server deployment not factoring in additional costs (e.g. monitoring costs, other software licenses, domain name costs, or management costs from the DC, etc.).
So only goal is to increase revenue. Their second offer post (July 23rd) shows even cheaper/lower priced plans (256 MB/512 MB RAM VPSes) at 3.5 and 5/year. For the sake of simplifying the problem, only real seen costs per unit (VPS) sold is the cost for the IP. Assuming 2.36 dollars per IP/year (50 dollars a month per /24, 254 usable IPs per /24), the cheapest plan (256MB) of the current offer nets $1.14/year in "profit".
With updated numbers, 0.5 dollars per IP per year suggests a $3 dollar/year income for the cheapest priced plan.
Let's say everyone from here on out buys the 5 dollar/year plan, as the higher resources plan has reached it's "peak" after being on the front page for so long. After all, an ad is less effective the longer it just sits there.
They need to sell around 536 VPSes at the 5 dollar/year mark (not including additional IP costs you need to take on to provided dedicated IPv4) to break even. Actually factoring in additional IP costs will increase that VPS number. That will bring up the number of VPSes on a SINGLE HOST NODE to an estimated 912 VPSes. (This section no longer accurate after direct allocation numbers were given)
If they decide to purchase a new host node and migrate customers over, then their expenses also doubles. You can probably negotiate a better deal to reduce your per-unit-cost, but even then you'll probably have 200+ VPSes per node with the current projections.
WHAT DOES THIS RELATE TO THIS THREAD
This is pretty much it.
Be realistic. If you (the customer) want to give away free money and have a "who has more servers idling" contest then go for it. But if you're expecting this thread to be taken seriously and really want to engage in a conversation on resources commit vs price point, then be realistic with your requests. Different companies have different targets on what they deem as a success. You want your provider to be sustainable so that they're able to continue to provide you those resources and plans long-term. You don't want to keep moving server-to-server and place-to-place. You have to recognize that sometimes getting a sustainable offer means your own time is valued more and you don't have to worry about "Is this updated and maintained?"
EDIT: This post was updated to reflect the new data Boom directly released publicly over Discord. The numbers shows that BoomerHost is profitable. However, I'll still add that this assumes a $0 in compensation from Boom's own time investment, taxes, and other un-calculated values.
However, my recommendation is still be realistic with what you expect out of your provider. I'd like for the people I work with to be compensated for their time or else they might:
In my opinion/perspective, you shouldn't be comparing the products outlined here as comparable to something like AWS/GCP/Azure or Linode/DigitalOcean/etc. But rather using the resources others are willing to leave behind to subsidize the operation/costs for the customers who actually use the resources. I believe XVMLabs was also experimenting with this methodology.