Yet they call everything they do a Cloud. Not only containers hooked up to EBS.
Cloud VPS is really cloud?
- Thread starter kunnu
- Start date
You will find very less true cloud service providers. Just labelling the cloud services to most common known services terms will give the boost to the business.
As majority of the world still does not know the true meaning of "cloud". The cloud vps services you are mentioning which are labelled with such low prices are just marketing gimmicks to attract the newbie clients.
As majority of the world still does not know the true meaning of "cloud". The cloud vps services you are mentioning which are labelled with such low prices are just marketing gimmicks to attract the newbie clients.
Licensecart
Active Member
That's the same with Unlimited everything hosts, they are marketing gimmicks as Joe said above, people think they will get it or that their server / hosting will never go down.
A cloud infrastructure should support cloud applications. Cloud applications are designed to be fault tolerant, and a cloud application operator would not want the decreased disk speed and increased pricing that comes with distributing virtual machine disk contents across multiple locations. Instead, operators primarily want to be able to provision and delete instances on the fly to scale up/down the application depending on usage, and also to be able to snapshot instances and upload their own disk images for provisioning.There are some shops that are really close to what a cloud should be:
1- Essentially is all about redundancy
2- and location independent.
If the node crash and the vm continue running without downtime, that is well.. cloud...
If a tsunami hit the DC and the vm continue running from another location without downtime ,, that is .. well.. that cant be more cloud
But real "cloud" if that even exist, we could say extreme redundancy" , doesn't cost 10 dollars per GB RAM and not 20, or 30. It really cost a lot.
Just a thought...
The idea of making individual virtual machines highly available is possible, but it is in some sense opposite to the idea of "cloud". With "cloud", we expect the application itself to be replicated in multiple locations and scalable so that it can handle changing numbers of users and changing user geographic location; but a single VM being highly available doesn't really provide that. So really, your idea would be targeting people running low end services where they don't have enough time to design a fault tolerant application, and instead want a cheap solution to get a highly available VM.
Note: the technology does exist, either by using a distributed storage across datacenters or, better yet, using something like Remus in Xen for primary-backup replication. Of course, you're going to need a dedicated link between the datacenters to handle that bandwidth, so they're going to have to be nearby.
Yes, people can build cloud applications without EBS. Besides, demanding that they have a message saying "yo, this part of our infrastructure is cloud, but this part isn't" is simply ridiculous; people can use any cloud infrastructure without taking advantage of the cloud capabilities, e.g. by only provisioning a single instance.
Last edited by a moderator:
Short of the obvious analogy to rain... um, the concept of 'redundancy' is already a kind of leakage in and of itself. Especially to the extent that data is being mirrored beyond RAID and at multiple locations across the world. It is preemptively leaked in a sense.
Can you clarify what you disagree with in my reply above? I argued that a cloud infrastructure should support cloud applications; if we accept that, then the only requirement is horizontal (not vertical) scalability and multiple locations. Also it's unclear what you mean by "single hypervisor"; Digital Ocean runs hundreds if not thousands of physical servers.
The idea of an ultra-reliable individual virtual machine on distributed storage with automatic failover is a dream that is entirely orthogonal to the idea of cloud, where we implement redundancy and high availability at the application level across an arbitrarily large number of VMs.
Scalable to people who actually design cloud applications means you can provision and delete VMs on the fly, NOT resize a VM up/down. The expectation is that you build your website to run on multiple VMs so that it remains online even if one VM fails. Perhaps you should explain why you believe your definition is at all valid instead of uselessly stating it without comment.
That's the same with Unlimited everything hosts, they are marketing gimmicks as Joe said above, people think they will get it or that their server / hosting will never go down.
Cloud has nothing to do with the reliability of individual VMs. When you're operating a cloud application that scales over thousands of VMs, you expect a few to fail every day.
Last edited by a moderator:
drmike
100% Tier-1 Gogent
Redundancy and reliability go together. You can use a lot of unreliable cheap garbage then waste time and energy trying to make such redundant. Bad utilization of time, money and energy. Yes people are doing such in various ways. People also live on 59 cent sandwiches and pay $100 a month to live in a tent also. Hardly good to go cheap most times.
If we are creating a RAID of VPS instances, surely I expect to have a fail once every X months to X years. I build to survive whatever. However, I tolerate few failures. Crap providers get kicked to the curb ASAP.
Even if one cheaps out and goes like $7 VPS instances and has to survive multiple fails plus scale, you need N+1 = 3 and since failure rate so high maybe 6, possibly 10. That's a lot of plumbing and noise on packets flying all over. Working on creating your own self DDoS.
No sane person expects or builds for multiple fails every day. In computing that rate of failure would be INSANITY.
I not total agree with you. We found a way to had a real cloud platform (as we mean cloud) to be affordable as normal vps. The 'problem' is initial investementthat avoid summer hoster or little companes to launch new low end real cloud because it take time to see money come back. this is not total bad for customer that can lean on strongest companies.
Build real cloud and keep same normal vps price is possible, medium/big competitors can do it easily, but little guys need to change they view. I mean they should stop buying one or few normal server but need to ask for funds, create real infrastructure easily scalable then launch the product. This cut recurring costs and you need to be prepared to be in passive for some months.
Not easy, but possible...
Cloud means lots of things, shared storage, lots of servers and an orchestrator or a cloud management interface.
All combined together to provide companies with hardware independence, high availability, flexibility and scalability. And of course most cloud provider are not really selling cloud!
All combined together to provide companies with hardware independence, high availability, flexibility and scalability. And of course most cloud provider are not really selling cloud!
at this moment "the cloud" is just alternative name for VPS or VDS. We do beta testing of true cloud, where we use currently 3 data centres + 6 dedicated servers and all data is secured and redundant.
As in my opinion the cloud should have:
- redundant data
- the load spread across all nodes
- multiple geo location (better to access services from anywhere worldwide)
- 100% uptime & secure
As in my opinion the cloud should have:
- redundant data
- the load spread across all nodes
- multiple geo location (better to access services from anywhere worldwide)
- 100% uptime & secure
farhanideas
Member
Cloud vps will be called as cloud if you receive cloud user option which help to create vps for your client in same pannel