ServerBear Performance Comparison of Rackspace, Digital Ocean, Linode and Vultr
-
@Dashrender said:
@wrx7m said:
@scottalanmiller said:
Here is the blog response to that...
https://blog.linode.com/2016/01/29/christmas-ddos-retrospective/
This was really interesting.
Wow - this sounds nearly the same as the GRC DDOS attack, only on a HUGE scale.
Yeah, the scale here is crazy.
-
@scottalanmiller said:
@Dashrender said:
@wrx7m said:
@scottalanmiller said:
Here is the blog response to that...
https://blog.linode.com/2016/01/29/christmas-ddos-retrospective/
This was really interesting.
Wow - this sounds nearly the same as the GRC DDOS attack, only on a HUGE scale.
Yeah, the scale here is crazy.
80 GB attack - damn!
-
@scottalanmiller I got busy and never got back to this.
I just cancelled the pings. here are the results.
64 bytes from 172.99.75.133: icmp_seq=34832 ttl=50 time=39.6 ms ^C --- 172.99.75.133 ping statistics --- 34832 packets transmitted, 20406 received, 41% packet loss, time 34860502ms rtt min/avg/max/mdev = 38.082/39.183/122.033/2.152 ms [root@keygen ~]#64 bytes from 108.61.151.173: icmp_seq=34830 ttl=54 time=49.9 ms ^C --- 108.61.151.173 ping statistics --- 34830 packets transmitted, 34821 received, 0% packet loss, time 34873888ms rtt min/avg/max/mdev = 45.852/47.901/842.335/7.692 ms [root@keygen ~]#64 bytes from 104.236.119.59: icmp_seq=34838 ttl=56 time=48.6 ms ^C --- 104.236.119.59 ping statistics --- 34838 packets transmitted, 34807 received, 0% packet loss, time 34890326ms rtt min/avg/max/mdev = 48.341/48.750/1051.915/5.427 ms, pipe 2 [root@keygen ~]# -
-
@aaronstuder that would be the Atom processor. Per thread performance is low. You get a lot of memory with that but not much CPU. Great for a lab. Much slower thread performance on physical than the VMs we are testing.
-
Linode is 5001.3 bogomips x 2 = 10,002.6
Scaleway is 4787.8 bogomips x 8 = 38,302.4
The C2750 is 85% more energy efficient than the Intel Xeon E5-2680 CPU.
In a data center with thousands of servers, that adds up quickly
http://www.cpu-world.com/Compare/900/Intel_Atom_C2750_vs_Intel_Xeon_E5-2680.html
-
have you looked at something like packet.net
-
@aaronstuder said:
Linode is 5001.3 bogomips x 2 = 10,002.6
Scaleway is 4787.8 bogomips x 8 = 38,302.4
You are comparing two cores to a full eight on a dedicated CPU. That the Linode is greater than 1/4 the bogomips when virtualized shows the difference.
And while yes, energy efficiency is important, so is largest scalable size and thread performance. Atoms are decent for certain workloads, but you need more of them to do it. But workloads for which Atoms are good, ARMs are better. So Atom often falls into a weird middle ground of "mostly only well suited for something that they are not ideal for."
-
@larsen161 said:
have you looked at something like packet.net
Decent looking product, but no RAID on their entry level. Really meant for DevOps only.
-
@scottalanmiller They have RAID.....
-
@aaronstuder said:
@scottalanmiller They have RAID.....
Not according to their price list. Only more expensive options begin to offer RAID.
-
@scottalanmiller said:
@aaronstuder said:
@scottalanmiller They have RAID.....
Not according to their price list. Only more expensive options begin to offer RAID.
They are super pricey too... $300 a month, almost. Granted, that's their Type 1 system... I get half the CPU & RAM, but almost 10 x the storage for $25 a month through KimSufi.
-
@dafyre said:
I get half the CPU & RAM, but almost 10 x the storage for $25 a month through KimSufi.
Your still using that? How do you like it?
-
@aaronstuder said:
@dafyre said:
I get half the CPU & RAM, but almost 10 x the storage for $25 a month through KimSufi.
Your still using that? How do you like it?
It's pretty good! I tried running XS on it, but XS simply does not like having one public IP and NATing VMs behind it... So I'm running KVM on it now. It's great!
I use Duplicity to backup my VMs and then ship the backups off to my Amazon Cloud Drive. I do not think their systems have RAID, but I get 16GB RAM, a 4 Core CPU (I forget which one) and 2TB of Spinning Rust. Bandwidth is excellent too.
I don't have any major workloads on it... Just a Chat server for my pops and about 10 people, a testing web server, and my ownCloud server with ~3 users.
-
@dafyre Can you explain how to do it with KVM?
-
@scottalanmiller said:
We need some latency numbers from around the world. Anyone want to collect some for us?
Here is the first IP address. A long running ping (hundreds or thousands of pings) would be good, we need the final stats from that:
- 104.236.119.59
- 108.61.151.173
- 172.99.75.133
We have a good idea on bandwidth, IO, CPU and memory. Network latency is pretty huge.
1,200,000 packets later...
via rackspace 8 GB General Purpose v1 based in london--- 162.242.243.171 ping statistics --- 400510 packets transmitted, 400499 received, 0% packet loss, time 400849652ms rtt min/avg/max/mdev = 79.018/79.634/184.775/2.778 ms --- 104.236.119.59 ping statistics --- 400759 packets transmitted, 400732 received, 0% packet loss, time 401132556ms rtt min/avg/max/mdev = 73.663/74.533/191.571/2.203 ms --- 108.61.151.173 ping statistics --- 400765 packets transmitted, 400749 received, 0% packet loss, time 401117767ms rtt min/avg/max/mdev = 69.861/75.792/205.164/3.167 ms
