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
There is lots of room between the extremes of $3 and $20 per TB ...
@TheLinuxBug
Interesting thoughts but I disagree. Reason: resource management and smartly fitting in combinations are a solved problem (albeit in other industries, it seems). So it is perfectly possible to offer the flexibility ("Lego system") I talked about. The only problem I would see is large upscaling, e.g. from 1 core and 512 MB mem to 6 cores and 12 GB mem - but with todays cloud infrastructure that can be done, too, especially when upscaling means (behind the scenes) creating a new VPS (possibly on another node) and transfering the system from the old node to the new one.
Frankly, I think the major problem is the f_cking (stupid) panels almost all providers use. In fact, I'm convinced, quite many providers also lack know-how and simply ride the wave of whatever the major panels offer.
I suggest in automating DA license, the customer orders a monthly VPS instead of flexible billing. That way it keeps things separate and clear, and also adding the option to pay for the VPS as prepaid monthly for the above arrangement, instead of having to top up $10 to use hourly.
This works for me.
And if someone is going to spend so much time setting up DA and configure it - only to delete the VPS, then they would better worry about the wasted effort and not the $ 3 USD (or higher as their plan may be).
On a side note- when time permits: adding the email response and the above explanation to the Knowledgebase would help :-)
https://intovps.com/client/index.php?rp=/knowledgebase
My other concern - bandwidth- has been discussed at length in the previous posts so I will leave it at that.
I like what I used so far, hope my mini review till see the light of the day soon.
You can disagree all you want. Again, I didn't say impossible or not eventual, however, I am explaining how, for now, the company I work for handles this. I have actually had it explained and demonstrated to me the importance of these metrics in maintaining a quality platform. That said, I am sure some small host where they are not as worried about service quality may be able to pull this off in some way. I don't know that I would suggest that model unless you have some ridiculously huge nodes where you can allocate weird amounts to some people and not have it effect the overall node in a negative way, though. Some of those running the Epyc servers with 250-500GB ram and NvME drives may be okay to do this, but someone using more conservative hardware it wouldn't be advisable.
my 2 cents.
Cheers!
It's not like "lego" cloud systems don't exist on the market.
But breaking apart a LET low cost package into logo pieces from which clients can pick and choose has no possible economic sense for a provider.
You get extraordinary value for your money from many LET providers.
Need a lego cloud system? You'll find it with the much higher priced providers.
That might work a lot better for continuity- $ 10 default advance payment for the base plan anyways works out to a little over 3 months. Or 1.5 months for the $6/month.
For the base plan,
If the VPS + DA setup works for the first month, a user has 2 more months to take a call for the Medium to LT perspective.
Discounting the teething troubles if any for the first month, if things do not work out in months 2 or 3, they can simply cancel and move on.
Anyways, I am sure the OP has thought out the pricing plans well and each user knows their constraints better.
Different things. One is a "prepackaged" special offer VPS and the other is a more general concept (and yes, the special offer could well be a part of it).
My point with you is
That said, I have no ill will whatsoever wrt @intovps and I apologize if you feel that I'm derailing your offer thread.
Until of course a customer orders 50 256mb VMs with 10gb disk (of which he makes 5gb swap) with 16 cpu cores each and both CPU and disk performance is shot, with most available ram unallocated.
Most providers carefully calculate what resources to allocate to get the most out of a system. You'll pay a premium if there isn't that balance.
I understand and agree. Probably some form of balance is required to avoid extremes.
That's exactly the reason you cannot chose VMs like 20 CPU cores and 1GB RAM.
CPU and RAM are too tightly connected to have whatever you want. Of course a provider can have two or more hardware configs, one with lots of RAM, one with lots of cores, and create the illusion that you pay only for what you need. Guess what? You're fully paying that slice of the server and price per cors and per those 256MB RAM increments will reflect that.
However, you can do that with storage if it's network attached. Obviously you need redundant, reliable, low latency, wide b/w connection to the storages and that's what we'ee doing in Cluj.
You're picturing a world where the LET bottom cost packages would live happily with a pick and chose system.
So you "complain" that the +3$ upgrade gives way too much, you only want +1TB of traffic, not the +2GB RAM, not the +10GB SSD NVMe. And you're making the assumption that a provider can offer you just the +1TB traffic at let's sat $1 and ONLY IF you actually use it.
It's not working like that. One in a hundred customers will actually use the whole traffic, that's what makes the whole package that low priced. And as I was saying, responsible overselling is good for everyone. But you take an oversold (in a good way) package, you want to pick and chose and pay for what you ACTUALLY USE, which means no overselling.
Right. Not going to happen. Take at look at traffic pricing of the big clouds. You pay only for what you use there.
Really no problem.
I'm also just trying to debunk a myth and enjoy the conversation.
@intovps
Kindly note that I'm not insane nor do I hunt weird dreams.
Yes, I want a better traffic price. But NO, I do not expect to pay it only if fully used. What I'm taking about is the option to upgrade only e.g. the traffic for a given package or only the memory. Whether and how much of it I use it is my problem, not yours. As for the burst traffic option I meant something that is not so uncommon, e.g. with "hot" periods of businesses where they have considerably more traffic. Of course I would not expect such a burst option to carry the same price tag as a standard package. It seems natural to me that such a capability would cost something on top (of the normal traffic price).
And please can we stop that annoying premise of the oh so friendly santa claus providers? You providers, at least in 95+% of cases, do not run special offers because you are such nice people. You do it for reasons like attracting new customers or like earning at least some money on some unused equipment, etc.
Well noted, that's perfectly fine with me. What is not fine however is to paint it as if you acted out of generosity. You make your living with us customers and we both need each other and accordingly we should treat each other well and honestly.
This is where the "serverless" (aka - pay for what resources you need and resources will be allocated as you need them) model is effectively moving to.
serverless... ba ha ha ha HA . and ha
I'll probably eat my words in a few years though.
"serverless" is a really stupid name. FaaS (Function as a Service) is much better.
@jsg I don't think Santa does overselling. I don't think I am Santa.
Just explaining to you why what you're looking for does not exist.
Actually "lego pieces" is a really illustrative name. What's the Lego price compared with the raw material? That's how Lego clouds are priced as well.
I wish you guys had some RCS&RDS Direct peering/IP Transit for Cluj location.
We have 2 fibers from RCS & RDS in Bucharest.
Why does it route via UPC.ro and/or via CZ then?
I've tested some ips from your looking glass a few minutes ago.
In Cluj we don't have direct RCS & RDS connection, but in our Bucharest location.
Bucharest, Romania:
Test IP address: 93.113.174.139
Do you have a looking glass for Bucharest location? Thank you.
Nope. We'll install one next week. hosterion.ro is hosted in Bucharest.
Congrats @intovps for showing Romanian love... have a few clients in Cluj that can benefit from the location. Atleast a better decent provider, the other one here on LET is a crack
Thank you
Nice rack, what model are those switches?
Edge-core AS6712-32X, 32 x QSFP+ ports of 40 GbE and running Cumulus Linux.
https://www.edge-core.com/_upload/images/AS6712-32X_ONIE_DS_R03_0829.pdf
A bit different than a regular LET setup
Hi, I certainly agree that $10 deposit is completely normal - what I was pointing out is that there was no mention of that upfront unless you happened to read the FAQ. And don't be ridiculous, nobody would seriously expect to prepay $0.0045! At least I hope not.
To be clear, it is not the fact that a $10 deposit is required, or that service is prepaid. It was just a surprise to me at 4 in the morning! I have been lucky perhaps (?) in that all my services are monthly billed in advance for the exact amount, or postpaid.
@intovps has certainly earned my respect (and future business) by their willingness to lower the initial deposit to $7 or even $3 on request. I won't be making such a request, but will be signing up soonish.
We've been charging a $10 credit ever since we've launched the OpentStack cloud, more than 2 years ago. I've overlooked to mention it in the initial post.
We stand by our offer and by the LET rules. So PM me after you place your order.
This look nice! I guess I should try to recover my account from 2010 and see what changed, it was nice back then but routing from home was meh and I had to cancel :<
EDIT: Nope, security question from 2010 (if I ever set that tbh ) stops me from password resetting
@intovps Looking at your pictures, I'm not seeing where your 1G public uplinks are, nor do I see any fiber channel switches. The 40G switches you've linked are regular Ethernet switches.
Contact support. They should be able to help you.