EPYC Basically Dedicated 16G, NEW FEATURE and more (RYZEN, XEON & MORE). Open Sourcing Software too!

eb1995 · June 2025

@SilverCreek said:

@eb1995 said:

@eb1995 said:

@SilverCreek said:

@Nvl said:
Please help me I have paid but Even cancelled This is the invoice number Invoice #109183 And now I haven't received any reply on the open ticket

Hi,

If you placed an order, but did not pay within a certain amount of time, the invoice is automatically cancelled so stock can be returned for others to order. You had opened this invoice on 06/09/2025, it's also important to note you placed 3 orders all of which were flagged for fraud.

I note that you opened 3 tickets for the orders, of which all have been responded to within our response time targets. #7695041: replied to 2 days ago. #1411681, #1541227: replied to 22 hours ago. I notice that you have not answered these tickets since our response.

@eb1995 said:

That’s interesting, this is the only provider I remember cancelling a limited promo order as I went to go to pay for it. I just don’t get why they cancelled, and because it was a so unexpected it’s hard to forget.

I encourage you to open a ticket or send me a message, as we generally don't cancel invoices. If it was during BF/CM or Christmas we did up the cancel timers for unpaid invoices as people were ordering, taking the stock, keeping their invoice unpaid for long periods of time.

@jsg said: Everything quite vague.
Q: What's a basically dedicated? A: Kind of you know ...
Kindly provide something tangible like processor model, fair share means what, does basically dedicated at (the very) least mean "100% usage, all the time"?
Also you make it sound like "we already have many locations - but now even more will come!" when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.
Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.
That of course might change in case you decide to finally provide some tangible and concrete information like e.g. a map or list of currently actually operational locations, clear VPS specs, and what basically dedicated actually is and means.

Basically Dedicated are virtual servers which are given higher priority over our standard VPS lineup. This includes higher I/O priority, memory, and CPU. You can reasonably use the entirety of the resources (and many do). We don't specifically dedicate the CPU by pinning it as it can get quite messy in our experience, a long time ago we offered VDS in the traditional sense and we had a lot of nightmares with pinning CPUs. That said, with our Basically Dedicated lineup, we've had no issues with clients using their resources and a lot do. We only had a few clients we had to ask to leave (crypto miners) based on our Terms of Service.

Basically Dedicated runs on various CPU models including EPYC 7402P, 7502, 7552, 7742. All underlying nodes sit with a minimum of 384GB of memory, our highest density nodes having just shy of 1.2TB of memory and dual 7742's. Our average CPU utilization remains under 75% per node, memory utilization remains under 80% per node, and we always have an excess of 4 TB worth of storage per node (which we add to as needed).

when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.

Unfortunately, we put stock in and we end up selling out quite quickly. On LET, we only end up putting ~30-40 in stock per post, of which you can select any location. We do similar on other communities or channels we advertise on, putting various products or discounts in stock. We also cap the coupon code usage, and each advertising channel gets it's own unique coupon so we can see where we get the most traction, and adjust our stock on the next post based on last sale's ingress traction (click -> order -> pay -> complete).

We often are ordering new nodes from our vendors, racking, installing. At the same time we're often ripping out old hardware (Xeon-grade equipment) to make space for more equipment. As an example, last weekend we pulled out 4 old E3 SuperMicros, 6 Dell R630's, an older SuperMicro E5 blade enclosure and replaced it all with EPYC, Ryzen units which are headed into production next week.

Our standard operating procedure for new nodes is racking them, benching them, memtesting, running various long-running stability tests, maximizing the CPU on and off for 1-2 weeks (ensuring airflow, no thermal throttling), and then putting it in stock. Our last additional to the lineup (High Performance), we had sitting in our cabinets across CA, US, EU for 96 days before we put it into production. This was not because there was a fault with the hardware (being it was 96 days), but we had to iron down the specifications, pricing on it as well.

Suffice it to say, we always have a significant amount of spare hardware, it's also why more often than not clients will be live-migrated across nodes, the underlying node has preemptive maintenance performed (eg, adding more memory, upgrading CPU, etc), and re-migrated back. This happens with nobody seeing any impact from a migration for preemptive maintenance to be performed.

a map or list of currently actually operational locations

I think you'll find the below list more exhaustive, including size of POP, purpose.

Canada:
- Montreal, QC: Operational (Cage)
- Ottawa, ON: Operational (Multi-Cab)
- Toronto, ON: Edge (Single Cab, Network Only)
- Vancouver, BC: Mixed (Single Cab, Limited Compute, Enterprise Only clientele), Edge (Network)

United States:
- New York, NY: Edge (Network Only)
- Dallas, TX: Operational (Multi-Cab)
- North Kansas City, MO: Operational (Single Cab)
- Seattle, WA: Edge (solely for SIX, backhaul to NKC)
- Los Angeles, CA: Edge (Network Only)
- Miami, FL: In Negotiation
- North Carolina, NC: Negotiated, Not Operational Yet (July 1st, 2025 becomes operational)

Europe:
- London, UK: Operational (Multi-Cab)
- Amsterdam, NL: Operational (Single, POP being expanded)
- Czechia, CZ: In Negotiation (TTC Teleport, if we go to Czechia)

Asia:
- Singapore, SG: No longer a POP (we exited SG back in late 2024, plan to re-enter SG around December, 2025)
- Auckland, NZ: Operational (Single)
- Sydney, AU: Mixed (Single, Limited Compute, Enterprise Only clientele at this time), Edge (Network)

In Asia, we rely almost exclusively on DataCamp/CDN77 for IP transit. As a side note, we recently entered into an agreement to bring DataCamp's transit and volume to all our facilities across Canada, US, Europe except a few (e.g., Ottawa due to lack of presence by DataCamp).

Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.

No problem, I'm always happy to answer questions. On our next post, I'll update the FAQ section with some of these answers as well to make it clearer.

I hope this helped clear up your questions.

Edit: Quote formatting.
Edit 2: Forgot to mention NC.

Ok thanks I sent a ticket. It was apparently marked as fraud but I was in the uk and never get marked as fraud

I got a reply but all it did it point out the obvious. Is that it?

We actually see this one was marked as Fraud on order. This would be why your invoice was cancelled.
0a12e11c510c7fbe9bbf90226dde39da62505b4ec0b19920e7a22470fd643357663a948aec68489c?t=38285b588940e6fc29f9596c9a948221

Thanks,
Curtis

Your ticket contents was to say you never get marked as fraud, and we responded showing you that it did get marked by the system, then said that is why the invoice was cancelled.

I just looked at your ticket history with us, and you never opened a ticket after your order was marked as fraud. It would've simply been marked as fraud automatically on order (score was 67/100 via FraudLabs and 41/100 with our internal fraud metric, anything >50 on either FL or internal is auto-marked for manual review on order), stock automatically returned to the system.

No, my first ticket was of a screenshot of the service where it has the details and it’s marked as fraud. And then a comment saying I never get marked as fraud I.e. a false positive. I did not just say I never get marked as fraud so you had to point out I did.

But good luck with that attitude.

jsg · June 2025

@Motion3549 said:

@jsg said:

@eb1995 said:

@eb1995 said:

@SilverCreek said:

@Nvl said:
Please help me I have paid but Even cancelled This is the invoice number Invoice #109183 And now I haven't received any reply on the open ticket

Hi,

If you placed an order, but did not pay within a certain amount of time, the invoice is automatically cancelled so stock can be returned for others to order. You had opened this invoice on 06/09/2025, it's also important to note you placed 3 orders all of which were flagged for fraud.

I note that you opened 3 tickets for the orders, of which all have been responded to within our response time targets. #7695041: replied to 2 days ago. #1411681, #1541227: replied to 22 hours ago. I notice that you have not answered these tickets since our response.

@eb1995 said:

That’s interesting, this is the only provider I remember cancelling a limited promo order as I went to go to pay for it. I just don’t get why they cancelled, and because it was a so unexpected it’s hard to forget.

I encourage you to open a ticket or send me a message, as we generally don't cancel invoices. If it was during BF/CM or Christmas we did up the cancel timers for unpaid invoices as people were ordering, taking the stock, keeping their invoice unpaid for long periods of time.

@jsg said: Everything quite vague.
Q: What's a basically dedicated? A: Kind of you know ...
Kindly provide something tangible like processor model, fair share means what, does basically dedicated at (the very) least mean "100% usage, all the time"?
Also you make it sound like "we already have many locations - but now even more will come!" when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.
Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.
That of course might change in case you decide to finally provide some tangible and concrete information like e.g. a map or list of currently actually operational locations, clear VPS specs, and what basically dedicated actually is and means.

Basically Dedicated are virtual servers which are given higher priority over our standard VPS lineup. This includes higher I/O priority, memory, and CPU. You can reasonably use the entirety of the resources (and many do). We don't specifically dedicate the CPU by pinning it as it can get quite messy in our experience, a long time ago we offered VDS in the traditional sense and we had a lot of nightmares with pinning CPUs. That said, with our Basically Dedicated lineup, we've had no issues with clients using their resources and a lot do. We only had a few clients we had to ask to leave (crypto miners) based on our Terms of Service.

Basically Dedicated runs on various CPU models including EPYC 7402P, 7502, 7552, 7742. All underlying nodes sit with a minimum of 384GB of memory, our highest density nodes having just shy of 1.2TB of memory and dual 7742's. Our average CPU utilization remains under 75% per node, memory utilization remains under 80% per node, and we always have an excess of 4 TB worth of storage per node (which we add to as needed).

when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.

Unfortunately, we put stock in and we end up selling out quite quickly. On LET, we only end up putting ~30-40 in stock per post, of which you can select any location. We do similar on other communities or channels we advertise on, putting various products or discounts in stock. We also cap the coupon code usage, and each advertising channel gets it's own unique coupon so we can see where we get the most traction, and adjust our stock on the next post based on last sale's ingress traction (click -> order -> pay -> complete).

We often are ordering new nodes from our vendors, racking, installing. At the same time we're often ripping out old hardware (Xeon-grade equipment) to make space for more equipment. As an example, last weekend we pulled out 4 old E3 SuperMicros, 6 Dell R630's, an older SuperMicro E5 blade enclosure and replaced it all with EPYC, Ryzen units which are headed into production next week.

Our standard operating procedure for new nodes is racking them, benching them, memtesting, running various long-running stability tests, maximizing the CPU on and off for 1-2 weeks (ensuring airflow, no thermal throttling), and then putting it in stock. Our last additional to the lineup (High Performance), we had sitting in our cabinets across CA, US, EU for 96 days before we put it into production. This was not because there was a fault with the hardware (being it was 96 days), but we had to iron down the specifications, pricing on it as well.

Suffice it to say, we always have a significant amount of spare hardware, it's also why more often than not clients will be live-migrated across nodes, the underlying node has preemptive maintenance performed (eg, adding more memory, upgrading CPU, etc), and re-migrated back. This happens with nobody seeing any impact from a migration for preemptive maintenance to be performed.

a map or list of currently actually operational locations

I think you'll find the below list more exhaustive, including size of POP, purpose.

Canada:
- Montreal, QC: Operational (Cage)
- Ottawa, ON: Operational (Multi-Cab)
- Toronto, ON: Edge (Single Cab, Network Only)
- Vancouver, BC: Mixed (Single Cab, Limited Compute, Enterprise Only clientele), Edge (Network)

United States:
- New York, NY: Edge (Network Only)
- Dallas, TX: Operational (Multi-Cab)
- North Kansas City, MO: Operational (Single Cab)
- Seattle, WA: Edge (solely for SIX, backhaul to NKC)
- Los Angeles, CA: Edge (Network Only)
- Miami, FL: In Negotiation
- North Carolina, NC: Negotiated, Not Operational Yet (July 1st, 2025 becomes operational)

Europe:
- London, UK: Operational (Multi-Cab)
- Amsterdam, NL: Operational (Single, POP being expanded)
- Czechia, CZ: In Negotiation (TTC Teleport, if we go to Czechia)

Asia:
- Singapore, SG: No longer a POP (we exited SG back in late 2024, plan to re-enter SG around December, 2025)
- Auckland, NZ: Operational (Single)
- Sydney, AU: Mixed (Single, Limited Compute, Enterprise Only clientele at this time), Edge (Network)

In Asia, we rely almost exclusively on DataCamp/CDN77 for IP transit. As a side note, we recently entered into an agreement to bring DataCamp's transit and volume to all our facilities across Canada, US, Europe except a few (e.g., Ottawa due to lack of presence by DataCamp).

Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.

No problem, I'm always happy to answer questions. On our next post, I'll update the FAQ section with some of these answers as well to make it clearer.

I hope this helped clear up your questions.

Edit: Quote formatting.
Edit 2: Forgot to mention NC.

Ok thanks I sent a ticket. It was apparently marked as fraud but I was in the uk and never get marked as fraud

I got a reply but all it did it point out the obvious. Is that it?

We actually see this one was marked as Fraud on order. This would be why your invoice was cancelled.
0a12e11c510c7fbe9bbf90226dde39da62505b4ec0b19920e7a22470fd643357663a948aec68489c?t=38285b588940e6fc29f9596c9a948221

Thanks,
Curtis

FWIW: I don't think that they are evil or intentionally ignorant, I think they simply aren't capable to see questions from the other's perspective and to respond to questions clearly and actually helpfully. In fact, I'm confident that @SilverCreek sincerely wants to answer the question but simply fails to get what we mean and want to know.
Plus, maybe there's an "inner lawyer" in his mind, insisting that nothing that might be taken against them is written. Frankly, his "answers" remind me of the way lawyers talk ...

The solution for us is simple: One either vibes with that - or not, in which case one stays away (like myself).

I'm not sure why you make hard pressure to someone telling they provide vs someone who does not telling their overselling

???

@SilverCreek said:

@jsg said:

@eb1995 said:

@eb1995 said:

@SilverCreek said:

@Nvl said:
Please help me I have paid but Even cancelled This is the invoice number Invoice #109183 And now I haven't received any reply on the open ticket

Hi,

If you placed an order, but did not pay within a certain amount of time, the invoice is automatically cancelled so stock can be returned for others to order. You had opened this invoice on 06/09/2025, it's also important to note you placed 3 orders all of which were flagged for fraud.

I note that you opened 3 tickets for the orders, of which all have been responded to within our response time targets. #7695041: replied to 2 days ago. #1411681, #1541227: replied to 22 hours ago. I notice that you have not answered these tickets since our response.

@eb1995 said:

That’s interesting, this is the only provider I remember cancelling a limited promo order as I went to go to pay for it. I just don’t get why they cancelled, and because it was a so unexpected it’s hard to forget.

I encourage you to open a ticket or send me a message, as we generally don't cancel invoices. If it was during BF/CM or Christmas we did up the cancel timers for unpaid invoices as people were ordering, taking the stock, keeping their invoice unpaid for long periods of time.

@jsg said: Everything quite vague.
Q: What's a basically dedicated? A: Kind of you know ...
Kindly provide something tangible like processor model, fair share means what, does basically dedicated at (the very) least mean "100% usage, all the time"?
Also you make it sound like "we already have many locations - but now even more will come!" when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.
Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.
That of course might change in case you decide to finally provide some tangible and concrete information like e.g. a map or list of currently actually operational locations, clear VPS specs, and what basically dedicated actually is and means.

Basically Dedicated are virtual servers which are given higher priority over our standard VPS lineup. This includes higher I/O priority, memory, and CPU. You can reasonably use the entirety of the resources (and many do). We don't specifically dedicate the CPU by pinning it as it can get quite messy in our experience, a long time ago we offered VDS in the traditional sense and we had a lot of nightmares with pinning CPUs. That said, with our Basically Dedicated lineup, we've had no issues with clients using their resources and a lot do. We only had a few clients we had to ask to leave (crypto miners) based on our Terms of Service.

Basically Dedicated runs on various CPU models including EPYC 7402P, 7502, 7552, 7742. All underlying nodes sit with a minimum of 384GB of memory, our highest density nodes having just shy of 1.2TB of memory and dual 7742's. Our average CPU utilization remains under 75% per node, memory utilization remains under 80% per node, and we always have an excess of 4 TB worth of storage per node (which we add to as needed).

when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.

Unfortunately, we put stock in and we end up selling out quite quickly. On LET, we only end up putting ~30-40 in stock per post, of which you can select any location. We do similar on other communities or channels we advertise on, putting various products or discounts in stock. We also cap the coupon code usage, and each advertising channel gets it's own unique coupon so we can see where we get the most traction, and adjust our stock on the next post based on last sale's ingress traction (click -> order -> pay -> complete).

We often are ordering new nodes from our vendors, racking, installing. At the same time we're often ripping out old hardware (Xeon-grade equipment) to make space for more equipment. As an example, last weekend we pulled out 4 old E3 SuperMicros, 6 Dell R630's, an older SuperMicro E5 blade enclosure and replaced it all with EPYC, Ryzen units which are headed into production next week.

Our standard operating procedure for new nodes is racking them, benching them, memtesting, running various long-running stability tests, maximizing the CPU on and off for 1-2 weeks (ensuring airflow, no thermal throttling), and then putting it in stock. Our last additional to the lineup (High Performance), we had sitting in our cabinets across CA, US, EU for 96 days before we put it into production. This was not because there was a fault with the hardware (being it was 96 days), but we had to iron down the specifications, pricing on it as well.

Suffice it to say, we always have a significant amount of spare hardware, it's also why more often than not clients will be live-migrated across nodes, the underlying node has preemptive maintenance performed (eg, adding more memory, upgrading CPU, etc), and re-migrated back. This happens with nobody seeing any impact from a migration for preemptive maintenance to be performed.

a map or list of currently actually operational locations

I think you'll find the below list more exhaustive, including size of POP, purpose.

Canada:
- Montreal, QC: Operational (Cage)
- Ottawa, ON: Operational (Multi-Cab)
- Toronto, ON: Edge (Single Cab, Network Only)
- Vancouver, BC: Mixed (Single Cab, Limited Compute, Enterprise Only clientele), Edge (Network)

United States:
- New York, NY: Edge (Network Only)
- Dallas, TX: Operational (Multi-Cab)
- North Kansas City, MO: Operational (Single Cab)
- Seattle, WA: Edge (solely for SIX, backhaul to NKC)
- Los Angeles, CA: Edge (Network Only)
- Miami, FL: In Negotiation
- North Carolina, NC: Negotiated, Not Operational Yet (July 1st, 2025 becomes operational)

Europe:
- London, UK: Operational (Multi-Cab)
- Amsterdam, NL: Operational (Single, POP being expanded)
- Czechia, CZ: In Negotiation (TTC Teleport, if we go to Czechia)

Asia:
- Singapore, SG: No longer a POP (we exited SG back in late 2024, plan to re-enter SG around December, 2025)
- Auckland, NZ: Operational (Single)
- Sydney, AU: Mixed (Single, Limited Compute, Enterprise Only clientele at this time), Edge (Network)

In Asia, we rely almost exclusively on DataCamp/CDN77 for IP transit. As a side note, we recently entered into an agreement to bring DataCamp's transit and volume to all our facilities across Canada, US, Europe except a few (e.g., Ottawa due to lack of presence by DataCamp).

Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.

No problem, I'm always happy to answer questions. On our next post, I'll update the FAQ section with some of these answers as well to make it clearer.

I hope this helped clear up your questions.

Edit: Quote formatting.
Edit 2: Forgot to mention NC.

Ok thanks I sent a ticket. It was apparently marked as fraud but I was in the uk and never get marked as fraud

I got a reply but all it did it point out the obvious. Is that it?

We actually see this one was marked as Fraud on order. This would be why your invoice was cancelled.
0a12e11c510c7fbe9bbf90226dde39da62505b4ec0b19920e7a22470fd643357663a948aec68489c?t=38285b588940e6fc29f9596c9a948221

Thanks,
Curtis

FWIW: I don't think that they are evil or intentionally ignorant, I think they simply aren't capable to see questions from the other's perspective and to respond to questions clearly and actually helpfully. In fact, I'm confident that @SilverCreek sincerely wants to answer the question but simply fails to get what we mean and want to know.
Plus, maybe there's an "inner lawyer" in his mind, insisting that nothing that might be taken against them is written. Frankly, his "answers" remind me of the way lawyers talk ...

The solution for us is simple: One either vibes with that - or not, in which case one stays away (like myself).

I'm not sure if we are getting each other. I believe I answered your questions as you wrote them, if you don't believe I did, can you re-state them in a clearer way? It's possible we just aren't understanding the true meaning of the question you're asking.

I guess that's basically what I wrote. And a decent start ...
(but sorry, I've lost my interest)

SilverCreek · June 2025

@JabJab said:
@SilverCreek you need better people skills.
They don't ask Why the invoice was cancelled, he is asking what he can do to get it working aka I want my server dude.

Thanks for the feedback.

In his ticket he didn't mention that they want to re-instate it, instead they just mentioned that it was marked as fraud and said they never get marked as fraud. I personally don't like to jump to conclusions but we're also talking about an order that was before we launched Amsterdam (order placed 26/12/2024). If it was a more recent order, I would have assumed that they want to have it re-instated.

siemens · June 2025

@SilverCreek said:

@JabJab said:
@SilverCreek you need better people skills.
They don't ask Why the invoice was cancelled, he is asking what he can do to get it working aka I want my server dude.

Thanks for the feedback.

In his ticket he didn't mention that they want to re-instate it, instead they just mentioned that it was marked as fraud and said they never get marked as fraud. I personally don't like to jump to conclusions but we're also talking about an order that was before we launched Amsterdam. If it was a more recent order, I would have assumed that they want to have it re-instated.

Will you have stock available soon?

SilverCreek · June 2025

@eb1995 said:

@SilverCreek said:

@eb1995 said:

@eb1995 said:

@SilverCreek said:

@Nvl said:
Please help me I have paid but Even cancelled This is the invoice number Invoice #109183 And now I haven't received any reply on the open ticket

Hi,

If you placed an order, but did not pay within a certain amount of time, the invoice is automatically cancelled so stock can be returned for others to order. You had opened this invoice on 06/09/2025, it's also important to note you placed 3 orders all of which were flagged for fraud.

I note that you opened 3 tickets for the orders, of which all have been responded to within our response time targets. #7695041: replied to 2 days ago. #1411681, #1541227: replied to 22 hours ago. I notice that you have not answered these tickets since our response.

@eb1995 said:

That’s interesting, this is the only provider I remember cancelling a limited promo order as I went to go to pay for it. I just don’t get why they cancelled, and because it was a so unexpected it’s hard to forget.

I encourage you to open a ticket or send me a message, as we generally don't cancel invoices. If it was during BF/CM or Christmas we did up the cancel timers for unpaid invoices as people were ordering, taking the stock, keeping their invoice unpaid for long periods of time.

@jsg said: Everything quite vague.
Q: What's a basically dedicated? A: Kind of you know ...
Kindly provide something tangible like processor model, fair share means what, does basically dedicated at (the very) least mean "100% usage, all the time"?
Also you make it sound like "we already have many locations - but now even more will come!" when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.
Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.
That of course might change in case you decide to finally provide some tangible and concrete information like e.g. a map or list of currently actually operational locations, clear VPS specs, and what basically dedicated actually is and means.

Basically Dedicated are virtual servers which are given higher priority over our standard VPS lineup. This includes higher I/O priority, memory, and CPU. You can reasonably use the entirety of the resources (and many do). We don't specifically dedicate the CPU by pinning it as it can get quite messy in our experience, a long time ago we offered VDS in the traditional sense and we had a lot of nightmares with pinning CPUs. That said, with our Basically Dedicated lineup, we've had no issues with clients using their resources and a lot do. We only had a few clients we had to ask to leave (crypto miners) based on our Terms of Service.

Basically Dedicated runs on various CPU models including EPYC 7402P, 7502, 7552, 7742. All underlying nodes sit with a minimum of 384GB of memory, our highest density nodes having just shy of 1.2TB of memory and dual 7742's. Our average CPU utilization remains under 75% per node, memory utilization remains under 80% per node, and we always have an excess of 4 TB worth of storage per node (which we add to as needed).

when actually you seem to have only a few locations operational and (all in all) hardly one rack full of nodes ... most of which are sold out.

Unfortunately, we put stock in and we end up selling out quite quickly. On LET, we only end up putting ~30-40 in stock per post, of which you can select any location. We do similar on other communities or channels we advertise on, putting various products or discounts in stock. We also cap the coupon code usage, and each advertising channel gets it's own unique coupon so we can see where we get the most traction, and adjust our stock on the next post based on last sale's ingress traction (click -> order -> pay -> complete).

We often are ordering new nodes from our vendors, racking, installing. At the same time we're often ripping out old hardware (Xeon-grade equipment) to make space for more equipment. As an example, last weekend we pulled out 4 old E3 SuperMicros, 6 Dell R630's, an older SuperMicro E5 blade enclosure and replaced it all with EPYC, Ryzen units which are headed into production next week.

Our standard operating procedure for new nodes is racking them, benching them, memtesting, running various long-running stability tests, maximizing the CPU on and off for 1-2 weeks (ensuring airflow, no thermal throttling), and then putting it in stock. Our last additional to the lineup (High Performance), we had sitting in our cabinets across CA, US, EU for 96 days before we put it into production. This was not because there was a fault with the hardware (being it was 96 days), but we had to iron down the specifications, pricing on it as well.

Suffice it to say, we always have a significant amount of spare hardware, it's also why more often than not clients will be live-migrated across nodes, the underlying node has preemptive maintenance performed (eg, adding more memory, upgrading CPU, etc), and re-migrated back. This happens with nobody seeing any impact from a migration for preemptive maintenance to be performed.

a map or list of currently actually operational locations

I think you'll find the below list more exhaustive, including size of POP, purpose.

Canada:
- Montreal, QC: Operational (Cage)
- Ottawa, ON: Operational (Multi-Cab)
- Toronto, ON: Edge (Single Cab, Network Only)
- Vancouver, BC: Mixed (Single Cab, Limited Compute, Enterprise Only clientele), Edge (Network)

United States:
- New York, NY: Edge (Network Only)
- Dallas, TX: Operational (Multi-Cab)
- North Kansas City, MO: Operational (Single Cab)
- Seattle, WA: Edge (solely for SIX, backhaul to NKC)
- Los Angeles, CA: Edge (Network Only)
- Miami, FL: In Negotiation
- North Carolina, NC: Negotiated, Not Operational Yet (July 1st, 2025 becomes operational)

Europe:
- London, UK: Operational (Multi-Cab)
- Amsterdam, NL: Operational (Single, POP being expanded)
- Czechia, CZ: In Negotiation (TTC Teleport, if we go to Czechia)

Asia:
- Singapore, SG: No longer a POP (we exited SG back in late 2024, plan to re-enter SG around December, 2025)
- Auckland, NZ: Operational (Single)
- Sydney, AU: Mixed (Single, Limited Compute, Enterprise Only clientele at this time), Edge (Network)

In Asia, we rely almost exclusively on DataCamp/CDN77 for IP transit. As a side note, we recently entered into an agreement to bring DataCamp's transit and volume to all our facilities across Canada, US, Europe except a few (e.g., Ottawa due to lack of presence by DataCamp).

Apologies if I got something wrong, but I feel like seeing quite a few signs telling me to stay away.

No problem, I'm always happy to answer questions. On our next post, I'll update the FAQ section with some of these answers as well to make it clearer.

I hope this helped clear up your questions.

Edit: Quote formatting.
Edit 2: Forgot to mention NC.

Ok thanks I sent a ticket. It was apparently marked as fraud but I was in the uk and never get marked as fraud

I got a reply but all it did it point out the obvious. Is that it?

We actually see this one was marked as Fraud on order. This would be why your invoice was cancelled.
0a12e11c510c7fbe9bbf90226dde39da62505b4ec0b19920e7a22470fd643357663a948aec68489c?t=38285b588940e6fc29f9596c9a948221

Thanks,
Curtis

Your ticket contents was to say you never get marked as fraud, and we responded showing you that it did get marked by the system, then said that is why the invoice was cancelled.

I just looked at your ticket history with us, and you never opened a ticket after your order was marked as fraud. It would've simply been marked as fraud automatically on order (score was 67/100 via FraudLabs and 41/100 with our internal fraud metric, anything >50 on either FL or internal is auto-marked for manual review on order), stock automatically returned to the system.

No, my first ticket was of a screenshot of the service where it has the details and it’s marked as fraud. And then a comment saying I never get marked as fraud I.e. a false positive. I did not just say I never get marked as fraud so you had to point out I did.

But good luck with that attitude.

If I misunderstood something, please accept my apologies.

allthemtings · June 2025

SilverCreek · June 2025

@siemens said:

@SilverCreek said:

@JabJab said:
@SilverCreek you need better people skills.
They don't ask Why the invoice was cancelled, he is asking what he can do to get it working aka I want my server dude.

Thanks for the feedback.

In his ticket he didn't mention that they want to re-instate it, instead they just mentioned that it was marked as fraud and said they never get marked as fraud. I personally don't like to jump to conclusions but we're also talking about an order that was before we launched Amsterdam. If it was a more recent order, I would have assumed that they want to have it re-instated.

Will you have stock available soon?

We can refill the Basically Dedicated 16G stock this evening. If you would like one, feel free to open a live chat or ticket and we can process one for you.

candle · June 2025

@SilverCreek Any chance for a similar offer on basically dedicated Ryzen 7900 like last time?
Would love to try singapore vps plans once it's available, and I'm assuming India/Mumbai POP is axed for now (?)

SilverCreek · June 2025

@candle said:
@SilverCreek Any chance for a similar offer on basically dedicated Ryzen 7900 like last time?
Would love to try singapore vps plans once it's available, and I'm assuming India/Mumbai POP is axed for now (?)

We have no stock on that right now, otherwise I would've included it. We underestimated the potential demand for it. I have another batch of hardware on order, but I don't expect we'll see it until next week, then in-line with our test procedure it's about 3-4 weeks before they're back in stock. Mumbai is in very early stages of negotiation.

siemens · June 2025

@SilverCreek said:

@siemens said:

@SilverCreek said:

@JabJab said:
@SilverCreek you need better people skills.
They don't ask Why the invoice was cancelled, he is asking what he can do to get it working aka I want my server dude.

Thanks for the feedback.

In his ticket he didn't mention that they want to re-instate it, instead they just mentioned that it was marked as fraud and said they never get marked as fraud. I personally don't like to jump to conclusions but we're also talking about an order that was before we launched Amsterdam. If it was a more recent order, I would have assumed that they want to have it re-instated.

Will you have stock available soon?

We can refill the Basically Dedicated 16G stock this evening. If you would like one, feel free to open a live chat or ticket and we can process one for you.

Ok, waiting on it.

bozolover99 · June 2025

Dang I bookmark to watch the other topic for when you restock and you started a new topic, missed out on the basically dedicated Montreal again...

siemens · June 2025

@bozolover99 said:
Dang I bookmark to watch the other topic for when you restock and you started a new topic, missed out on the basically dedicated Montreal again...

He said there will be stock yesterday, so probably soon you might get it.

bigbenn15 · June 2025

Basically Dedicated 32 coupon

bozolover99 · June 2025

@bigbenn15 said:
Basically Dedicated 32 coupon

Is there a coupon for this? It is more than I need

bigbenn15 · June 2025

i dont no this is good

NetPIMP · June 2025

@SilverCreek said:

@NetPIMP said:
If I already have "basically dedicated" vps's, can they be moved onto AMD-based nodes? (they're on xeon nodes at the moment)

Yes, if you happen to get provisioned on one of the legacy nodes we can move you over.

Fun Fact: opened the ticket 3 weeks ago... took them 2 weeks to determine that it's "doable" and a week after that, still sitting here... wondering if this is the level of service I should expect when renewal time comes around... if the answer was "we can't migrate you, sorry" then just say that...

SilverCreek · June 2025

@NetPIMP said:

@SilverCreek said:

@NetPIMP said:
If I already have "basically dedicated" vps's, can they be moved onto AMD-based nodes? (they're on xeon nodes at the moment)

Yes, if you happen to get provisioned on one of the legacy nodes we can move you over.

Fun Fact: opened the ticket 3 weeks ago... took them 2 weeks to determine that it's "doable" and a week after that, still sitting here... wondering if this is the level of service I should expect when renewal time comes around... if the answer was "we can't migrate you, sorry" then just say that...

Hey there, my apologies -- could you send me your ticket number in a DM? We have a few migrations sitting in our queue depending on what location you're in as we're deploying more hardware there currently. Shoot it over in a DM, I'll take a look shortly.

NetPIMP · June 2025

@SilverCreek said:

@NetPIMP said:

@SilverCreek said:

@NetPIMP said:
If I already have "basically dedicated" vps's, can they be moved onto AMD-based nodes? (they're on xeon nodes at the moment)

Yes, if you happen to get provisioned on one of the legacy nodes we can move you over.

Fun Fact: opened the ticket 3 weeks ago... took them 2 weeks to determine that it's "doable" and a week after that, still sitting here... wondering if this is the level of service I should expect when renewal time comes around... if the answer was "we can't migrate you, sorry" then just say that...

Hey there, my apologies -- could you send me your ticket number in a DM? We have a few migrations sitting in our queue depending on what location you're in as we're deploying more hardware there currently. Shoot it over in a DM, I'll take a look shortly.

sent ticket# via DM 7 days ago - I think it was within a few minutes of your reply ... literally zero feedback... 🤷🏻‍♂️

ourvds · July 2025

Open source sounds interesting)

Howdy, Stranger!

Categories

In this Discussion

EPYC Basically Dedicated 16G, NEW FEATURE and more (RYZEN, XEON & MORE). Open Sourcing Software too!

Comments

Howdy, Stranger!

Quick Links

Categories

In this Discussion

EPYC Basically Dedicated 16G, NEW FEATURE and more (RYZEN, XEON & MORE). Open Sourcing Software too!

Comments