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
This would be interesting.
@SilverCreek coupon doesn't work for 32GB
Can someone please explain VDS vs this "basically dedicated" vps?
Can't submit ticket if it doesn't work
what do you mean by Basically Dedicated?
is it similar to VDS?
also curious about this, seems like one of the better deals i've seen here, the one im on rnow is 8/16 160 from lch for 7.50, and I might swap here for higher storage, have a bunch of little selfhosted apps on there and it fills up fast.
My apologies for that, we were in the middle of overnight maintenance to update department settings, ingress methods and alerting. It should be good now!
It is materially similar, just our spin on it.
was sceptical at the start of this 'basically dedicated' but you seem to have a loyal following, GLWS!
Sceptical in what way?
it's not dedicated, and I suspect some users would get confused. already some in the thread asking exactly what it means.
>
Hola! Is SYD supposed to be live?
Everything seems to be the three US locations.
But the resources are dedicated, right?
@SilverCreek what is standard vs premium transit? do you have a looking glass for testing?
they are but they aren't, that's the rub. I'll leave to provider to explain
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
No stock
@SilverCreek
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.
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.
Yeah, weird.
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.
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.
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).
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.
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).
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
Sorry but that feels just like the usual vague sales blabla. I'm not interested in which processor models can/may/might drive the VPS and VDS(?) - I want to know which model a node that you sell in a promo really is driven by.
Just as I want to get a clear and concrete definition of what "basically dedicated" means.
So, does it or does it not at least mean "can be used 100% all the time, no ifs or buts, period"?
We don't need yet more new words for 'VDS', we need to know what we actually get.
And thank you for your informative albeit somewhat basic location list! I appreciate that.
I believe our AUP outlines quite nicely (alongside the TOS) permissible usages (https://terabit.io/aup).
Speaking generally, we reserve the right to ask you to upgrade to a higher specification plan if your usage is constantly hitting the barriers of the plan you're on.
To date, outside of people running prohibited activities (crypto miners, zmaping the planet), we have never had to ask someone to leave or upgrade. We have also not had to enforce any other policies related to the service as permitted by our Terms and Acceptable Usage.
We have many clients fully consuming their storage (our highest being two single VMs consuming 800GB of NVMe total), CPU and memory allocated to them.
To be absolutely clear, you can utilize all the resources allocated to you as long as the utilization is in compliance with our policies. We branded this as Basically Dedicated to distinct from the fact that we are not doing resource pinning for the reasons I had explained before. It does not mean you can't consume all the resources.
To clarify our migration policy:
To highlight #3 specifically, if we have a group of users who wants to do all their Plex transcoding (for example) or crunch some ML models, if they let us know in advance and space permits, we'll migrate them to a node with higher performance CPUs (3Ghz+ base clock), sometimes a GPU (we have 2 of these nodes in CA, US with a Tesla V100) depending on space and their use case. It's noteworthy that we don't charge any extra for this, and it's something we generally don't talk about publicly. If the request for us to be more extensive is there in terms of outlining our internal policies (beyond the TOS, AUP), I'd be happy to share some of our internal policies once it's approved internally specifically related to systems, facilities and related operations, and publicize them on our Knowledge Base.
I got a reply but all it did it point out the obvious. Is that it?
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
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.
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.
@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.