All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
DigitalOcean vs Linode vs Vultr
Hi,
I just got the below machine from the above mentioned host
1) DigitalOcean
VPS Type : kvm
CPU type : Intel(R) Xeon(R) CPU E5-2650L v3 @ 1.80GHz
Number of cores : 4
CPU frequency : 1799.998 MHz
RAM : 7823 MB
Swap : MB
I/O speed : 678MB/s
2) Linode
VPS Type : kvm
CPU type : Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Number of cores : 6
CPU frequency : 2499.996 MHz
RAM : 12012 MB
Swap : MB
I/O speed : 642MB/s
3) Vultr
VPS Type : kvm
CPU type : Virtual CPU a7769a6388d5
Number of cores : 6
CPU frequency : 2399.996 MHz
RAM : 7822 MB
Swap : MB
I/O speed : 288MB/s
And also I have ran the benchmarks
Left - Linode Right - DigitalOcean
Middle - Vultr
CPU - http://prnt.sc/dcxds6
Memory Read - http://prnt.sc/dcxg9a
Memory Write - http://prnt.sc/dcxhic
IO - http://prnt.sc/dcxixt
DB - http://prnt.sc/dcxqaa
Can you please tell me based on the CPU and Benchmark which is good to go with ?
And there is a chance I may be getting the below specs too
VPS Type : kvm
CPU type : Intel(R) Xeon(R) CPU E5-2697 v4 @ 2.30GHz
Number of cores : 6
CPU frequency : 2299.996 MHz
RAM : 12012 MB
- Which one to go with ?109 votes
- DigitalOcean22.94%
- Linode34.86%
- Vultr28.44%
- Newer CPU13.76%
Comments
Ok, wait a minute while the community comes back since they are all in an order queue purchasing cheap VPS deadpooling a few.
voted Vultr as I'm using their entry-level instance over DO but I honestly think you should shop around carefully before you commit $80 /mo on any of these three providers ..
Generally higher = better. Then again the benchmarks are completely meaningless unless they are measuring what you plan to do with said box(es).
To be honest all of these look so close to each other that I see no clear path to pick one over the other based on those screenshots. I'd say go with who you like based on other factors then. Price and support, perhaps.
Voted DO though because, well, reasons :P
If the network speed and ping is important for you, then try test it?
didn't even looked the specs, hitted that DO radio button and clicked submit.
I stayed with Linode for some years chiefly because their support is usually nice. You can max out the CPU usage without being chased by sysadmins. DO and vultr are also pretty stable, but I would not put important stuff on it because, you know, reasons.
DO. Vultr had weird TOS/AUP terms relatively recently (there are posts that can be found in LET/WHT archives, with samples).
However, to those who don't care about "fine print" and possible privacy/etc problems, Vultr offers more features than DO and definitely surpasses it in some aspects.
Personally, I choose DO and will stay well clear of Vultr. JMNSHO.
DO is premium stuff. Rarely fail and support is fast. I like how stable the VMs are. Stability is prime.
Also their documentation DO come handy.
Fast Support? Their Support is damn slow, at least by my standards...I have opened a few tickets and depending on the time of day it takes them hours.
Working on it
Worth noting though that it's still fast by the standard of other self-managed providers at that price point with more than a few thousand customers. Perspective is important. But that doesn't mean it isn't something to work on. Just important that the comparisons are not hosts that haven't met scale or managed providers.
Providing support is easy. Providing support at scale is a different task. Keeping support amazing as we grow, that's part of my job
You do know that those host aren't "True" cloud providers, imho. Ask them what raid they use, how many HD's data is being written on.
I love Linode, stable and much features
How many HDs makes one "true" cloud?
https://en.wikipedia.org/wiki/Cloud_computing
http://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-145.pdf
(FYI, this is the definition hosts are using when they say cloud, so if your definition varies, you cannot be an informed buyer, and your expectation should not be used to educate buyers or you will be teaching them to shop without proper expectations)
Another day, another thread brother. When I have the energy we can see who has more inches. :P
Everyone defines cloud differently and I can argue both sides but IMHO, they aren't true cloud providers.
Not the hosts, not the people writing about it in tech articles. If people selling cloud mean it to be one thing and you teach others that it means another, you are misinforming people and creating in them expectations that will cause them to purchase services with the expectation that they are buying something else. If people buy from a cloud provider and expect something else it's because of people misinforming them, not because of false advertising.
It's not about inches. It's about using not misusing your knowledge to mislead consumers just because you want the market to accept a different norm. It may very well please you to know that a few people sign up for cloud providers and complain because they've been misinformed, but it does not please me.
why not ramnode or dediserve?
Really? The people selling Cloud services can't even agree on what true Cloud is! I stated an opinion and for whats its worth even included "IMHO". This is a public form and every thread is half full of opinions. I just stated a popular opinion amongst many in our hosting communities/circles.
You aren't a true cloud provider in my book unless you have:
Failover Setup/Clustered Storage Architecture/High Redundancy
Having cloud like billing doesn't make you a true cloud, imho. (Hourly billing, scalability, etc)
Agreed. I just followed up with fact to ensure that any readers are well informed and have proper expectations. Because I care about potential customers of the entire industry, not just my own, and have no desire to see people misled in an effort to attempt to coerce market changes.
@jarland Random thoughts - I've always felt DO should run their own DNS recursion platform for their infrastructure to rely on. Linode do it, Vultr do it, I hate having to rely on Google.
I agree. That would be great. I've heard so many stories in the last year of Google rate limiting IP ranges, relying on their free resolvers does not seem like a solid end game for anyone.
Do you have IOPS benchmark?
I'm not sure what facts your presented and I can find plenty online to support my case but users here can do their own research and judge for themselves based on their needs. In all honesty I think you or anyone else is misleading ppl if you think the above mentioned companies are "True Cloud" providers. Honestly surprised to hear it coming from you since many in the industry above my pay grade and expertise level disagree with you.
We can agree to disagree. I was hesitant to even reply because this debate is waged every few months on various boards and sites but and no one wins but just as you felt it your responsibility to reply to my post and correct me, I felt the same towards you.
Let our posts be a springboard for future clients of Cloud services to research and decide for themselves if companies advertising themselves as cloud servers are really cloud.
https://en.wikipedia.org/wiki/Cloud_computing
http://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-145.pdf
https://www.ibm.com/cloud-computing/learn-more/what-is-a-cloud-server
I'm actually not sure what facts you presented. I'm not mad at you or anything, I do this for the readers. It saddens me every time I have to correct someone's expectation that was grounded not in the facts of what cloud providers sell and advertise, but the rumors that people spread on internet forums.
I care, so I inform. That is all there is to it.
I think you base that on forum conversation and not documented or scholarly information. I know that I base mine on documented scholarly information. You'll see why I have trouble seeing it from that perspective.
If everyone sells a yellow beach ball and you go around telling everyone that true beachballs are white, you are the reason that people are misinformed when they write a review on Amazon saying "Not a true beach ball, false advertising, stupid thing was yellow." If you're comfortable with that responsibility, that's not my concern I suppose. I, however, retain the right to step in to those forums and say "Actually, beach ball does not imply the color white."
I've done nothing to set those other expectations, and I do not wish to sell someone something that they believe is something else. Surely you would agree the desire to not get sales from misinformed customers is at least common decency. If others are setting different expectations on my behalf, I feel obligated to speak up. This in addition to the fact that I simply care about this industry as a whole, and informing buyers who wish to jump into this industry is important to me. I wouldn't want them buying from Amazon with the wrong expectation of cloud, assuming that a basic EC2 instance is distributed, any more than I would want them assuming it of a droplet.
Our stances on this matter are documented in this thread for future readers. I am not trying to debate you or engage in some online nerd, I.T. spar. I felt inclined to offer an opinion and I did.
Benchmarked then all and Linode wins especially now that majority of Vultr VPS are the lower clocked 2.4Ghz and not the 3.4+ Ghz ones so closer to the lower DO cpu performance.
some are older Linode XEN benchmarks back then so Linode KVM instances these days would be even faster
Newer cpu but not necessarily faster due to lower clock speed than Linode's E5-2680v3 at 2.5Ghz - depends on your work loads i.e. for encryption/decryption and https/ssl (especially for ECC 256bit ECDSA based SSL certificates), the E5-2697v4 might be faster or around the same as Linode E5-2680v3 just due to the difference in clock speeds of 2.3Ghz vs 2.5Ghz. E5 v4 would be faster clock for clock for encryption/decryption especially AES-NI based but the lower clock speed just might even things out compared to E5-2680v3.
Linode is also expanding it's network https://blog.linode.com/2016/11/02/network-update-multihomed-increased-transit-peering/.
FYI, I currently have VPS with Linode, Vultr and DigitalOcean
Just to point. This is not a grey area. It's defined within the National Institute of Standards and Technology, part of the US Department of Commerce. While definitions of words can change over time, the definition outlined by @jarland is currently the official accepted definition. Now you can say you define it differently sure, however issues could come up later during discussions. It's like saying climate change is just an opinion and you disagree with it. There will be people who agree with you, but when most of the scientific community have agreed upon the commonly accepted definition, saying otherwise could potentially be dangerous to those who are misinformed expecting more (and in the end, is absurd).
Many of the research labs I've worked with have accepted that the definition of cloud computing is what it is. Just using a server that's offsite connected through the internet. Now most universities do have their own computing clusters and hosting infrastructure so questions like "should we use AWS/Azure/GCP/DO?" isn't that common, but those situations do come up and while they don't expect their data to disappear one night, they accept that this is a risk.
Redundant data storage, while originally more traditional in earlier platforms, aren't necessarily a requirement to be defined as a cloud.
Note, other people can say "well that's just the US's official definition of the term" and they'll be correct (I haven't checked with any other local government's definitions). However, even in academia, many countries and researchers use the definitions outlined by the United States or Europe to provide the foundation for their own policies and regulations. It doesn't directly impact it, but rather indirectly influence that country's definition.
Use whatever definition you wish, but please be aware that there are standards to definitions. Misinforming readers who may not be well versed in these terminology could be very dangerous and could expose those users to risk they do not want to accept.
Edit: Those who believe "true cloud" as being something with high availability and redundancy, please quit being delusional. As a proper professional, communication is important in this industry and definitions are critical. Please do your research, but realize that most people who believe cloud includes high availability and redundancy don't have proper background in this technology or are mostly affiliated with some obscure entity. There's a reason NIST had to get involved in this matter. Having redundancy and high availability may be a feature involved (including clustering), but they are not de-facto of cloud hosting or computing.
P.S. This isn't an inches measuring contest. Nor is it a contest to begin with. It's a clarification of a definition.
And that's my post on here for the rest of the year.
Some DNS-based blacklist also rate limit using the recursive DNS resolver, if that gives DO another incentive
. Back when I was using DO for my email server I had to switch to HE.net's public DNS.
As the above benchmark has nothing to do with this post , but i want to share this stuff.
There is major performance increase in my website since I have switched from easyengine to VPSSIM
@eva2000 I decided to move away from easyengine and decided to give a try with VPSSIM and Centminmod , Installed both. But moved away with VPSSIM , since I don't know how to access your interface after installation. Is there any good guide for newbies ?
I have tried out easyengine and VPSSIM , VPSSIM is best with much option. I will be trying out Centminmod tonight
Thanks for all the contribution you people have made till now. And the Vultr is leading according to poll.
Done.