Hello Low End Spirit!
Pulsar67
Hosting Provider
Hi everyone,
I'm from Pulsar67. We’re currently working on building a new website focused on delivering quality cloud hosting services. At this stage, we’re still developing our infrastructure and refining our operations, so please hold off on any purchases—we’re not production ready yet but will have some offers for LES soon!
We currently operate services using colocation out of Tampa, Florida, and looking to expand! Let us know if you have any locations that we should also operate out of!
Non-Profit Partnerships:
We’re passionate about giving back to the community. As part of our mission, we’re looking to work with non-profits that support communities in any way. If you know of any organizations that could benefit from our services or that we should consider partnering with, please let us know. We’re eager to explore opportunities to make a positive impact together.
We truly value the feedback from this community. If you have any suggestions or notice any issues as we progress, please feel free to share. We’re here to learn, grow, and eventually serve you better.
Thanks for having us, and we look forward to engaging with you all!
Best regards,
The Pulsar67 Team
Comments
Checklist of quality cloud hosting service:
No hostname left!
Hello!
We're going to have IPv6—we have a /40 allocation. We're still determining the best allocation size for our customers, and we're trying to keep our setup versatile so that if someone needs a larger allocation, we can offer that.
We're definitely planning on adding the option to disable IPv4, possibly with savings too!
We currently don't support customer-sided rDNS updates. We will offer that functionality by opening a ticket; this is something we look into for the future.
We don't support client-ended ISO at the moment; however, this is something in the works, but I don't think it'll be ready before we start selling. We are able to mount ISOs if needed and will also include that in our FAQ while we work on that functionality.
CPU steal will be a metric we monitor once we start selling!
We don't have any throttling on the CPU. In the case of excessive CPU usage, we may ask you to upgrade or reach out to understand why. We don't have and won't implement any automated processes for throttling or suspending servers over this.
Definitely will look into some deals for LES!
Hah, It's so harsh
why not 1299?
The Ultimate Speedtest Script | Get Instant Alerts on new LES/LET deals | Cheap VPS Deals | VirMach Flash Sales Notifier
FREE KVM VPS - FreeVPS.org | FREE LXC VPS - MicroLXC
Just give everyone /64 like all other good providers do. You have enough of them.
The all seeing eye sees everything...
@Pulsar67
Hy and welcome, hope your project gets a good take-off
I recommend you 0, as anything above 0.1% will ultimately get you nowhere, read LES & OGF, that never worked for anyone.
I can agree on some level on this, it is really up to you how much $$$ you wish to burn for Marketing, if you go down this path just keep in mind that those are 100% loss
Even @yoursunny can agree that a /64 is enough, right??
, but yes, anything lower than that will cost you a ton of work to fix later ( it was for us )
Don't, at least do not focus on this right from the start, out of 100 IPV6 only services, 70% will ask for IPV4 within 24 hours, this is the truth for now, at least from our point of view.
That will work nicely for start
I would recommend you own your Routing and Switching hardware and do not rely on upstream or the Datacenter, but that is me. You can do whatever works for you.
Just my 2 cents on this, please do not:
use the hypervisor ( Proxmox, Xen, vMware and so on ) to do routing, use a Physical Router or a VM with a router OS installed ( vyos or any other )
do not use layer 3 switches as routers, they lack the ASIC's for filtering and high PPS Throughput, they also cannot handle 2 or more full BGP tables ( in case you will be multi-homed ). While some can do BGP that is true, that function was never intended to be used at ISP level, also some struggle with large IPV6 tables as they lack the RAM for it.
Do what you tested and use whatever you are comfortable waking up at 3 AM to fix ( this is something a friend of mine suggested to me, he was 100% right
)
Other then this, read the forums and see where others have failed, so you do not try to do the same thing.
Cheers and Godspeed !
HOST-C
Host-C - VPS & Storage VPS Services – Reliable, Scalable and Fast - AS211462
"If there is no struggle there is no progress"
FAILED
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
What happens if the customer ignores contacts while keeps consuming CPU?
We'd like it to be throttling instead of suspension.
Yes, /64 is enough for most servers during automated provisioning.
Options to request larger or additional prefixes can be offered via tickets.
One such use case is tunneling to a home router or another server that lacks sufficient IPv6 space.
We made such a tunnel from Cloudie YTZ to VirmAche EWR.
Since less-than-/64 equals headache, the tunnel needs its own /64, which in turn requires the source server to have more than /64.
Disabling IPv4 can be offered through pre-sales ticket.
We got this saving on our Salmon HKG server via ticket.
No hostname left!
The royal We. Crass!
It wisnae me! A big boy done it and ran away.
NVMe2G for life! until death (the end is nigh)
@yoursunny why not AS1299? What did Telia ever do wrong. I understand your concerns with Cogent.
I guess Wholesale should be fine as an upstream.
youtube.com/watch?v=k1BneeJTDcU
We would likely throttle the CPU if we don’t receive a response in a reasonable amount of time. We would only ever suspend a service for repeated abuse or extreme cases (ie: distributing malware, network abuse)
We will do /64 as that seems to be the general consensus for the best network size to offer. We will allow multiple or larger block sizes to be requested.
Still deciding on the whole v6 implementation.
Ignore the want to be trolls. Read, read, and read some more. Listen to what other hosts tell you about the market.
Free Hosting at YetiNode | MicroNode | Cryptid Security | URL Shortener | LaunchVPS | ExtraVM | Host-C | In the Node, or Out of the Loop?
Did a lot of research into what would be best for us and decided on going with Juniper for our routers & switches (both routers do L3 while both our switches only do Layer 2 switching). Been working very well for us in our test environment so far.
That's for sure definitely want to get everything right the first time around so it's smooth sailing and we aren't trying to make extreme changes on production.
Thank you for the suggestions!
1299 congested network.
No hostname left!
Bless the 3AM comment, that's true. Francisco told me something similar a couple years ago and it's true. Don't use or do something you're not willing to fix at 3AM.
For small networks (10g and below) building your own router is actually quite easy using off the shelf hardware and multi port cards. Not as good as buying a router, but it's a few hundred with opnsense instead of a few thousand for a cisco or juniper (if you're avoiding microtik / ubiquiti like the plague). Haven't experimented yet with 40g routing using home brew, that's up next on my list.
Find your niche - what can you do that others can't, you'll always excel better in what you can vs what others can't do. Use what you can to your advantage - but don't do so as a detriment to customers.
And the final tidbit of advice: Treat your customers well. Even if everything goes to shit, being respectful and honest to your client goes a long way towards not building resentment. I've had a fair share of mistakes along the way, and am not afraid to admit I've definitely done outright refunds for my mistakes.
Why limiting to 10g or below?
Mentally strong people build 100~200 Gbps router easy using off the Rhino hardware and multi port cards.
IPv4 + IPv6 + IPv9 all supported.
No hostname left!
Ban @yoursunny
Agree on this,
Jokes aside.
If you need good hardware at decent prices, @NathanCore4 can provide you with such, I can personally vouch for them, no issues for the past 1+ years of collaboration. Whatever we bought, price was
and the stuff was A+++
Host-C - VPS & Storage VPS Services – Reliable, Scalable and Fast - AS211462
"If there is no struggle there is no progress"
That looks like amateur reseller hosting. Who does rdns by ticket these days?
North Korea
Ratio
youtube.com/watch?v=k1BneeJTDcU
We will now have rDNS configurable on the client side when we launch.
Unless you will do MAIL services, rather focus on other more pressing stuff at go-live.
Host-C - VPS & Storage VPS Services – Reliable, Scalable and Fast - AS211462
"If there is no struggle there is no progress"
I’m almost ready to start a trial—just waiting on some hardware. In the meantime, I’m wrapping up anything else I can.
Everything is working and nearly production-ready, but I want to set up HA on my routers. I was originally planning to use Juniper for routing, but I’m now considering VyOS since, theoretically, I should get better performance for the same price.
Does anyone have experience with VyOS? I haven’t used it yet but am planning to spin up a VyOS instance later today.
vyos will be similar to junos ( similar not 1:1)
We tested it out, not bad at all, here is my 2 cents on software routing:
last resort
A dedicated device with specialized ASIC will always beat Software. Yes, that dedicated will have a higher price most of the time.
Host-C - VPS & Storage VPS Services – Reliable, Scalable and Fast - AS211462
"If there is no struggle there is no progress"
>
Any day of the week is this statement true
Free Hosting at YetiNode | MicroNode | Cryptid Security | URL Shortener | LaunchVPS | ExtraVM | Host-C | In the Node, or Out of the Loop?
@Pulsar67
If you wish to go down the path of Software Routing, please take these into consideration: ( these are based on what we tested and ultimately failed, but at least we learned something
)
@ 1 GBPS anything will do it perfect, regarding the OS or CPU even if it is a VM
@ 10 GBPS and above, things change, a lot, especially if you will do firewall, scrubbing or NAT ( the later being a killer latency wise )
a. using a x86 CPU will mean that most tasks will be single threaded, so high clock speed will always beat multi core in a server setup, 3.0 GHz would be a minimum, also you will need high memory bandwidth cpu generation, Xeon V2 will be a no-go )
b. bare metal beats vm, always, mostly stability wise
c. use only Ethernet cards that have good tcp hw offload, here chelsio is the winner, they specialize on this, T520-CR for 10 GBPS or T540-CR for 40 GBPS, whatever above that on a x86 will be a waste in setup and will start to cost as much as a dedicated device from juniper or cisco or other
d. RAM - will need to be as high frequency as possible, 32-64 gb will do just fine.
if you decide to go this way, maybe you can share your experience with us.
ASR1001-X 20GBPS licensed or ASR1002-HX ( comes 60 GBPS licensed by default as I recall ) are not that expensive nowadays on e-bay
Other providers can enlighten you on Juniper
But again, whatever you feel comfortable when shit hit's the fan and that will most probably happen at some point, it happens to all of us.
Cheers!
Host-C - VPS & Storage VPS Services – Reliable, Scalable and Fast - AS211462
"If there is no struggle there is no progress"
As others have stated, stay away from MikroTik as if it was the plauge.
VyOS is nice, but you need to spend som money to get good hardware (fast PCi buses and such).
Would recommend using a hardware router, Juniper MX204's are very nice (4x 100G), you can find MX104's cheap (they don't have the same capacity or density, but OK entry level).
You can also use a Arista 7280 (upgrade memory from default 8G to 32G), they have Flexroute and some other nice features, so you can run full BGP tables if required (not as fast as MX204's, but you get LOTS of port for much less money). Would recommend this over VyOS.
Arista switches are nice for ToR and aggregation switches.
We run a 100% Arista network with L3 leaf and spine design along with Juniper MX204s for edge routing.
Did look at many alternatives before going this route. MikroTik was one, due to their pricing, but the price doesn't justify the headaches it gives you using it.
Have been using Cisco routing gear since 97, since Cisco has been the only alternative, but nowadays, they are shit compared to the competition. Arista is like using Cisco, same CLI, migrating from Cisco to Arista was a dream. Juniper is another story, but getting used to it, also nice.
Just my long 2 cents.
Thank you everyone for your feedback! I've taken some time to review everything and have come up with a rough baseline of how I am going to deploy everything.
If we were to do anything in this route it would be bare metal. If I do decide to attempt VyOS I would definitely simulate traffic on the network before I set for production. Wouldn't like to figure out issues too late.
I would for sure share my experiences with it; however this first time around I am not doing a full rack which is why I think VyOS might work for me.. I would need to do some more research and testing before I come to that decision.
So far the Juniper switches I have purchased have been working very well for me I will definitely take a look at Arista in the future thank you for the suggestion!
Trust me and others on this, you cannot simulate whatever will 500 VPS or 1000 VPS customers will do, just Brace For Impact
.
You could assume that they will do mostly sub 900 byte packets and do the math from there, as that is almost true, SW OS Routing will suck at a high rate small packet PPS by the way, that is my personal issue on the matter.
It is always better to over power routing and switching performance, that will assure you have leg-room to modify stuff on the go.
MX204 is a fine choice to start with if you like Junos.
Arista is King latency wise and in switching, they kinda beat all the rest.
. Overkill if you will not do storage via NFS/ISCSI/CEPH or other storage services via Ethernet.
Focus more on customer stuff ( your products, what you bring to the table ), you can always upgrade devices on the go with announced downtime ( we did, oh boy did we change all in 2 years ), and sincerely, I do not think that anyone will mock you if you upgrade something on an announce downtime. ( there is always that 1% who will do
)
I did not say, as I assumed it is not the case; you do not want to throw 4x IPv4 /24 sub-nets on the same broadcast domain I hope.
Nor you wish to do that with IPv6.
Cheers!
PS:
You could also not bother with routing, and ask you upstream to publish your sub-nets and you just do switching - this is useful at start and if you have only 1 upstream, so you do not spend the $$ on routing and only do switching; it would lower your costs considerably on start.
Host-C - VPS & Storage VPS Services – Reliable, Scalable and Fast - AS211462
"If there is no struggle there is no progress"