Virtono Miami KVM Review

Note: the Cloud XL VPS used for this review was provided for free for benchmark purposes. This did not affect any of my conclusions in the thread. See the LET thread for more details of the giveaway they ran.

This is an initial review of the Virtono.com KVM cloud servers in Miami, based on under a week of usage.

Contents:
1. Pricing and Features
2. Benchmarks: Performance
2.a Serverbench.io
2.b Bench.sh
2.b Stability over time
2.c Network
3. Uptime
4. Security
5. Support/Communication
6. Conclusion

Pricing and Features

These are KVM VPS's with SSD storage. The Cloud-XL has, at the time of test, 8GB RAM, 120GB SSD, 4 cores (2.67GHz) and 8TB bandwidth, all for €43.95.

Virtono have run up to 30% discount recurring offers for quite some time in their recent LET posts.

Benchmarks: Performance**

Serverbench.io

Test results for CLOUD XL - MIAMI at Virtono

Server specs:

4 × Westmere E56xx/L56xx/X56xx (Nehalem-C)
8 GB RAM / 120 GB disk space
Ubuntu 18.04 Bionic
Miami, United States

Benchmark results summary:

UnixBench - 2761.3
Disk Read - 1825 MB/s
Disk Write - 1555 MB/s
Bandwidth - 845.88 MB/s

More: https://serverscope.io/trials/vVNx

Bench.sh

 CPU model            : Westmere E56xx/L56xx/X56xx (Nehalem-C)
 Number of cores      : 4
 CPU frequency        : 2666.760 MHz
 Total size of Disk   : 115.0 GB (1.9 GB Used)
 Total amount of Mem  : 7976 MB (90 MB Used)
 Total amount of Swap : 4095 MB (0 MB Used)
 System uptime        : 1 days, 19 hour 35 min
 Load average         : 0.02, 2.58, 4.15
 OS                   : Ubuntu 18.04.1 LTS
 Arch                 : x86_64 (64 Bit)
 Kernel               : 4.15.0-22-generic

 I/O speed(1st run)   : 570 MB/s
 I/O speed(2nd run)   : 602 MB/s
 I/O speed(3rd run)   : 447 MB/s
 Average I/O speed    : 539.7 MB/s

 Node Name                       IPv4 address            Download Speed
 CacheFly                        205.234.175.175         83.6MB/s      
 Linode, Tokyo, JP               106.187.96.148          6.21MB/s      
 Linode, Singapore, SG           139.162.23.4            4.09MB/s      
 Linode, London, UK              176.58.107.39           18.3MB/s      
 Linode, Frankfurt, DE           139.162.130.8           12.8MB/s      
 Linode, Fremont, CA             50.116.14.9             20.2MB/s      
 Softlayer, Dallas, TX           173.192.68.18           63.4MB/s      
 Softlayer, Seattle, WA          67.228.112.250          27.2MB/s      
 Softlayer, Frankfurt, DE        159.122.69.4            11.4MB/s      
 Softlayer, Singapore, SG        119.81.28.170           5.58MB/s      
 Softlayer, HongKong, CN         119.81.130.170          6.40MB/s      

 Node Name                       IPv6 address            Download Speed
 Linode, Atlanta, GA             2600:3c02::4b           80.8MB/s      
 Linode, Dallas, TX              2600:3c00::4b           36.2MB/s      
 Linode, Newark, NJ              2600:3c03::4b           29.0MB/s      
 Linode, Singapore, SG           2400:8901::4b           6.51MB/s      
 Linode, Tokyo, JP               2400:8900::4b           7.64MB/s      
 Softlayer, San Jose, CA         2607:f0d0:2601:2a::4    21.0MB/s      
 Softlayer, Washington, WA       2607:f0d0:3001:78::2    9.90MB/s      
 Softlayer, Paris, FR            2a03:8180:1301:8::4     11.6MB/s      
 Softlayer, Singapore, SG        2401:c900:1101:8::2     4.67MB/s      
 Softlayer, Tokyo, JP            2401:c900:1001:16::4    6.48MB/s      

Benchmarks: Stability

For the price of this plan, it's possible to get a very budget dedicated server. As a result, I was highly interested in the stability of this plan - or how noisy the node was - to determine whether I'd be better off getting a potentially lower, but more stable, performance elsewhere.

Over 3 days, from Wednesday to Friday, the following were run every hour:

• Ping latency tests (IPv4 and v6)

• 100MB download tests (IPv4 and v6)

• Sequential write tests (dd)

• Random IO tests (ioping)

I've found that these are useful tests to identify whether there is some periodic issue (e.g network congestion during peak hours).

Results

The main thing I was looking for was the absence of prolonged periods of abnormally slow speeds, which would indicate node/network congestion.

The results were fairly stable over time - it seems that any network deviations are fairly minor, and possibly due to external factors (such as a single speedtest server being slower for a short amount of time).

There were some periods of slow maximum IOPS (tested using ioping), but these did not last long. The sequential write speed was not affected during these times, so I'm not too concerned.

In total there are 72 readings, once each hour for 3 days.

Ping Latency (IPv4)

Ping Latency (IPv6)

Download Speed (IPv4)

Download Speed IPv6)

Disk Sequential Write Speed

Disk ioping results

Network

My IP was in a /24 announced by m247.com (as9009.net). M247 seem to have a diverse range of peers. The latencies and speeds in the speedtests I ran were all as good as I've seen elsewhere.

Uptime

During the 4 days before this review was written, the server had 100% uptime.

Security

According to spectre-meltdown-checker, mitigations for the Spectre/Meltdown CPU bugs are in place.

Support/Communication

I submitted a ticket on 25/09/2018 (20:42 local time for the provider) asking for an additional IPv6 address. This was provided - with a clear, helpful and friendly reply - within 14 minutes. A very impressive reply time given that it was a low priority ticket.

Conclusions

Although this VPS is outside the LET price range, it provides impressive and fairly stable performance. Support are fast and responsive. It will be interesting to see how it performs in the longer-term.

I didn't realise at first that this VM has "dedicated resources". This makes it a lot more appealing for knowing I can run CPU intensive programs without causing problems.