All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
xHosts promo review - Read!
I got a @xHosts promo VPS, mainly for location diversification (I like to have my servers a bit spread all over Europe plus one or two in NA and APAC and found the connectivity in Britland quite good), so I grabbed that promo VPS expecting halfway decent performance at the price.
Colour me shivered, I really didn't expect what I saw when doing a couple of benchmark runs (12).
Here you go, as usual, sys info, processor, and memory first
Version 2.5.0a, (c) 2018+ jsg (->lowendtalk.com)
Machine: amd64, Arch.: amd64, Model: AMD EPYC 7F52 16-Core Processor >!
OS, version: FreeBSD 14.2, Mem.: 989 MB
CPU - Cores: 1, Family/Model/Stepping: 23/49/0
Cache: 32K/32K L1d/L1i, 512K L2, 256M L3
Std. Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat
pse36 cflsh mmx fxsr sse sse2 sse3 pclmulqdq ssse3 fma cx16 sse4_1
sse4_2 popcnt aes xsave osxsave avx f16c rdrnd hypervisor
Ext. Flags: syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm lahf_lm cmp_legacy svm
cr8_legacy lzcnt sse4a misalignsse 3dnowprefetch osvw perfctr_core
AES? Yes
InNested Virt.? Yes
HW RNG? Yes
ProcMem SC [MB/s]: avg 267.3 - min 115.8 (43.3 %), max 418.0 (156.4 %)
ProcMem MA [MB/s]: avg 407.1 - min 402.0 (98.7 %), max 414.4 (101.8 %)
ProcMem MB [MB/s]: avg 408.5 - min 403.5 (98.8 %), max 412.6 (101.0 %)
ProcMem AES [MB/s]: avg 1325.5 - min 1319.3 (99.5 %), max 1335.1 (100.7 %)
ProcMem RSA [kp/s]: avg 119.9 - min 115.8 (96.6 %), max 124.5 (103.9 %)
Look at that! Not only very decent single and multi vCore performance but actually a tighter spread than my netcup VDS ("root server")! Plus very good crypto performance. Impressive!
Let's look at the drive performance
--- Disk 4 KB - Buffered ---
Write seq. [MB/s]: avg 7.75 - min 7.47 (96.4%), max 7.95 (102.6%)
Write rnd. [MB/s]: avg 7.08 - min 6.75 (95.3%), max 7.41 (104.6%)
Read seq. [MB/s]: avg 13.35 - min 12.88 (96.5%), max 13.82 (103.5%)
Read rnd. [MB/s]: avg 12.50 - min 10.73 (85.8%), max 14.17 (113.4%)
--- Disk 4 KB - Sync/Direct ---
Write seq. [MB/s]: avg 7.24 - min 6.92 (95.6%), max 7.41 (102.4%)
Write rnd. [MB/s]: avg 6.81 - min 6.65 (97.7%), max 6.98 (102.5%)
Read seq. [MB/s]: avg 13.40 - min 12.92 (96.4%), max 13.74 (102.5%)
Read rnd. [MB/s]: avg 12.64 - min 11.06 (87.5%), max 14.78 (116.9%)
--- Disk 64 KB - Buffered ---
Write seq. [MB/s]: avg 77.84 - min 73.99 (95.1%), max 79.94 (102.7%)
Write rnd. [MB/s]: avg 66.49 - min 64.02 (96.3%), max 69.39 (104.4%)
Read seq. [MB/s]: avg 4181.22 - min 3747.83 (89.6%), max 4672.77 (111.8%)
Read rnd. [MB/s]: avg 133.48 - min 125.81 (94.3%), max 140.59 (105.3%)
--- Disk 64 KB - Sync/Direct ---
Write seq. [MB/s]: avg 12.35 - min 9.46 (76.6%), max 17.06 (138.1%)
Write rnd. [MB/s]: avg 7.48 - min 5.02 (67.1%), max 14.48 (193.6%)
Read seq. [MB/s]: avg 4174.34 - min 3869.92 (92.7%), max 4512.70 (108.1%)
Read rnd. [MB/s]: avg 138.60 - min 125.82 (90.8%), max 151.67 (109.4%)
--- Disk 1 MB - Buffered ---
Write seq. [MB/s]: avg 102.46 - min 98.72 (96.3%), max 105.89 (103.3%)
Write rnd. [MB/s]: avg 185.96 - min 179.25 (96.4%), max 190.09 (102.2%)
Read seq. [MB/s]: avg 5213.10 - min 5035.27 (96.6%), max 5453.71 (104.6%)
Read rnd. [MB/s]: avg 649.24 - min 610.97 (94.1%), max 687.19 (105.8%)
--- Disk 1 MB - Sync/Direct ---
Write seq. [MB/s]: avg 59.45 - min 48.62 (81.8%), max 75.32 (126.7%)
Write rnd. [MB/s]: avg 74.67 - min 58.19 (77.9%), max 117.87 (157.9%)
Read seq. [MB/s]: avg 5299.78 - min 5085.29 (96.0%), max 5519.44 (104.1%)
Read rnd. [MB/s]: avg 640.37 - min 600.17 (93.7%), max 672.48 (105.0%)
--- Disk IOps (Sync/Direct) ---
Write seq. [MB/s]: avg 21.24 - min 19.67 (92.6%), max 21.98 (103.5%)
IOps : avg 5437.79 - min 5036.09 (92.6%), max 5625.98 (103.5%)
About 20 MB/s 4k4t and well over 5000 IOps. Very decent indeed and way beyond what I expected for less than €20 per year. That kind of performance is what I usually see at double that price and higher.
So far really impressive but surely there must be a weak spot, right? Probably mediocre connectivity. Let's see ...
--- Europe ---
NO OSL mirror.terrahost.no [F: 0]
DL [Mb/s]: avg 384.4 - min 356.7 (92.8%), max 409.1 (106.4%)
Ping [ms]: avg 28.1 - min 28.1 (99.9%), max 28.2 (100.2%)
Web ping [ms]: avg 28.7 - min 28.1 (97.9%), max 30.0 (104.6%)
UK LON lon.speedtest.clouvider.net [F: 0]
DL [Mb/s]: avg 3031.4 - min 2461.8 (81.2%), max 3427.9 (113.1%)
Ping [ms]: avg 2.1 - min 2.1 (99.2%), max 2.2 (103.9%)
Web ping [ms]: avg 2.2 - min 2.1 (93.7%), max 2.8 (124.9%)
NL AMS mirrors.xtom.nl [F: 0]
DL [Mb/s]: avg 1278.2 - min 1131.0 (88.5%), max 1496.0 (117.0%)
Ping [ms]: avg 7.6 - min 7.6 (100.0%), max 7.6 (100.0%)
Web ping [ms]: avg 8.6 - min 7.6 (88.6%), max 9.7 (113.1%)
DE FRA fra.lg.core-backbone.com [F: 0]
DL [Mb/s]: avg 884.7 - min 821.6 (92.9%), max 909.0 (102.7%)
Ping [ms]: avg 12.4 - min 12.3 (99.3%), max 12.6 (101.7%)
Web ping [ms]: avg 12.5 - min 12.3 (98.8%), max 12.9 (103.6%)
FR PAR mirror.in2p3.fr [F: 0]
DL [Mb/s]: avg 513.4 - min 371.1 (72.3%), max 571.5 (111.3%)
Ping [ms]: avg 18.9 - min 18.9 (100.0%), max 18.9 (100.0%)
Web ping [ms]: avg 49.3 - min 18.9 (38.3%), max 155.7 (315.9%)
CH GEN pkg.adfinis-on-exoscale.ch [F: 0]
DL [Mb/s]: avg 620.3 - min 584.8 (94.3%), max 642.0 (103.5%)
Ping [ms]: avg 18.1 - min 18.0 (99.6%), max 18.1 (100.1%)
Web ping [ms]: avg 18.3 - min 18.0 (98.2%), max 20.5 (111.9%)
IT MIL it1.mirror.vhosting-it.com [F: 0]
DL [Mb/s]: avg 479.0 - min 306.8 (64.1%), max 531.4 (110.9%)
Ping [ms]: avg 22.4 - min 22.3 (99.6%), max 22.5 (100.4%)
Web ping [ms]: avg 29.9 - min 22.3 (74.7%), max 60.3 (202.0%)
ES MAD mirror.es.stackscale.com [F: 0]
DL [Mb/s]: avg 404.1 - min 60.4 (14.9%), max 453.2 (112.2%)
Ping [ms]: avg 23.8 - min 23.7 (99.7%), max 23.9 (100.6%)
Web ping [ms]: avg 32.6 - min 24.9 (76.3%), max 56.7 (173.7%)
RO mirrors.hosterion.ro [F: 0]
DL [Mb/s]: avg 312.8 - min 303.4 (97.0%), max 319.9 (102.3%)
Ping [ms]: avg 36.1 - min 36.0 (99.7%), max 36.2 (100.3%)
Web ping [ms]: avg 36.1 - min 36.0 (99.7%), max 36.2 (100.3%)
GR UNK speedtest.ftp.otenet.gr [F: 12]
DL [Mb/s]: avg 0.0 - min 0.0 (0.0%), max 0.0 (0.0%)
Ping [ms]: avg 37.5 - min 0.0 (0.0%), max 45.1 (120.3%)
Web ping [ms]: avg 37.5 - min 0.0 (0.0%), max 45.1 (120.3%)
RU MOS speedtest.hostkey.ru [F: 0]
DL [Mb/s]: avg 245.1 - min 237.2 (96.8%), max 251.4 (102.6%)
Ping [ms]: avg 45.4 - min 45.3 (99.8%), max 45.5 (100.2%)
Web ping [ms]: avg 54.8 - min 45.4 (82.9%), max 61.0 (111.3%)
--- Asia / Oceania ---
RU SIB mirror.truenetwork.ru [F: 0]
DL [Mb/s]: avg 110.4 - min 106.7 (96.7%), max 112.8 (102.2%)
Ping [ms]: avg 102.0 - min 101.3 (99.3%), max 104.4 (102.3%)
Web ping [ms]: avg 104.3 - min 101.3 (97.1%), max 116.2 (111.4%)
IR TEH mirror.mobinhost.com [F: 0]
DL [Mb/s]: avg 128.4 - min 83.0 (64.6%), max 147.5 (114.9%)
Ping [ms]: avg 86.6 - min 85.9 (99.1%), max 87.1 (100.5%)
Web ping [ms]: avg 95.2 - min 86.1 (90.5%), max 108.9 (114.4%)
IN MUM mirrors.piconets.webwerks.in [F: 0]
DL [Mb/s]: avg 45.4 - min 42.2 (92.8%), max 48.0 (105.8%)
Ping [ms]: avg 249.7 - min 244.6 (97.9%), max 256.8 (102.8%)
Web ping [ms]: avg 254.0 - min 244.6 (96.3%), max 277.7 (109.3%)
SG SGP mirror.sg.gs [F: 0]
DL [Mb/s]: avg 70.3 - min 67.2 (95.6%), max 72.0 (102.4%)
Ping [ms]: avg 155.6 - min 155.5 (99.9%), max 155.8 (100.1%)
Web ping [ms]: avg 158.4 - min 157.0 (99.1%), max 160.8 (101.5%)
CN HKG mirrors.xtom.hk [F: 3]
DL [Mb/s]: avg 32.4 - min 0.0 (0.0%), max 49.2 (151.8%)
Ping [ms]: avg 262.7 - min 221.4 (84.3%), max 492.9 (187.6%)
Web ping [ms]: avg 265.5 - min 223.8 (84.3%), max 492.9 (185.7%)
CN NAJ mirror.nyist.edu.cn [F: 3]
DL [Mb/s]: avg 27.9 - min 0.0 (0.0%), max 47.4 (170.0%)
Ping [ms]: avg 275.0 - min 0.0 (0.0%), max 483.3 (175.8%)
Web ping [ms]: avg 280.2 - min 0.0 (0.0%), max 483.3 (172.5%)
CN BEJ mirrors.bfsu.edu.cn [F: 3]
DL [Mb/s]: avg 28.7 - min 0.0 (0.0%), max 46.0 (160.3%)
Ping [ms]: avg 194.2 - min 0.0 (0.0%), max 461.7 (237.7%)
Web ping [ms]: avg 228.2 - min 0.0 (0.0%), max 461.7 (202.3%)
JP TOK ftp.udx.icscoe.jp [F: 1]
DL [Mb/s]: avg 37.8 - min 0.0 (0.0%), max 49.9 (132.1%)
Ping [ms]: avg 228.3 - min 0.0 (0.0%), max 341.4 (149.5%)
Web ping [ms]: avg 242.9 - min 0.0 (0.0%), max 404.5 (166.5%)
AU SYD mirror.internet.asn.au [F: 0]
DL [Mb/s]: avg 46.7 - min 46.4 (99.4%), max 47.0 (100.8%)
Ping [ms]: avg 242.8 - min 242.7 (99.9%), max 243.3 (100.2%)
Web ping [ms]: avg 243.3 - min 242.7 (99.8%), max 245.0 (100.7%)
--- Africa ---
ZA JOB mirror.datakeepers.co.za [F: 0]
DL [Mb/s]: avg 70.2 - min 69.0 (98.4%), max 71.0 (101.2%)
Ping [ms]: avg 160.4 - min 160.2 (99.9%), max 161.1 (100.4%)
Web ping [ms]: avg 161.4 - min 160.9 (99.7%), max 161.8 (100.3%)
KE NAI mirror.liquidtelecom.com [F: 0]
DL [Mb/s]: avg 80.7 - min 74.9 (92.8%), max 85.4 (105.8%)
Ping [ms]: avg 145.2 - min 145.1 (99.9%), max 145.5 (100.2%)
Web ping [ms]: avg 145.5 - min 145.1 (99.8%), max 146.0 (100.4%)
--- America ---
US NYC nyc.mirrors.clouvider.net [F: 0]
DL [Mb/s]: avg 153.8 - min 144.4 (93.9%), max 158.3 (103.0%)
Ping [ms]: avg 71.6 - min 70.0 (97.8%), max 77.1 (107.7%)
Web ping [ms]: avg 71.7 - min 70.0 (97.6%), max 77.6 (108.2%)
US CHI ord.mirror.rackspace.com [F: 0]
DL [Mb/s]: avg 119.6 - min 55.9 (46.8%), max 133.5 (111.6%)
Ping [ms]: avg 90.5 - min 85.4 (94.4%), max 101.3 (111.9%)
Web ping [ms]: avg 91.8 - min 86.1 (93.8%), max 102.4 (111.6%)
US LAX mirror.alma.lax1.serverforge.org [F: 2]
DL [Mb/s]: avg 65.1 - min 0.0 (0.0%), max 84.8 (130.2%)
Ping [ms]: avg 200.6 - min 145.3 (72.4%), max 424.1 (211.4%)
Web ping [ms]: avg 200.6 - min 145.3 (72.4%), max 424.1 (211.4%)
Pardon me, are you kidding me, xHosts? 3 Gb/s to UK, LON I understand, after all, the server is in Britland and obviously quite near the target. But about 1 Gb/s, give or take 100 or 200 Mb/s to NL and DE is surprising. And even more surprising I find that the worst result within Europe is RO, BUC with still over 300 Mb/s (OK, Moscow is "only" about 250 Mb/s but then that's a quite different distance and story)!
Asia and Oceania is less impressive but still shows very good results. Both RU, Siberia and IR, Teheran well over 100 Mb/s, and Singapore about 70 Mb/s is quite decent. But yes, some targets show not exactly overwhelming results, India, Mumbai and China, Bejing and Nanjing are examples and so is Japan, Tokyo. But I've seen far worse, so "acceptable with some peaks" is my verdict. Oh, and Ozzyland has a quite decent result too.
Both african targets have decent results as well.
The us-american east coast results are really impressive. I don't seem to every have seen about 150 Mb/s to NYC and about 120 Mb/s for Chicago is really good as well. Only the west coast result is mediocre, but oh well, no surprise ...
TL;DR that's a hell of a deal!, kudos to xHosts, very well done! Highly recommended - with a potential but: maybe, just a suspicion, my VPS happens to be on a (yet) only lightly occupied node.
Comments
@jsg Thanks for the feedback, I have an appointment next week with my account manager and we normally any changes we would like to feedback.
We are always happy to hear feedback and look at how we can improve the service we offer.
On a side note, we are emailing full details out tonight, there are some ACI Routing upgrades planned within the next week.
Thanks for taking the time to order and write the review, it is greatly appreciated
An absolute bargain!
I have two Ryzen VPS with them and they have been great! They are easy to work with!
Oops, I've confused the two coasts, I apologize.
Maybe @angstrom or @Arkas could be so kind to edit the second to last paragraph and simply swap "east" (coast) and "west" (coast)?
Corrected as requested
8 billion people on this planet call that USA (or United States of America), I have no idea why you're always being a snowflake intentionally spelling things wrong.
Of course I purchased the @xHosts VPS with 140 TB monthly traffic volume (after commenting the order number)!
And of even bloody courser I benchmarked it right away. After all, the recently purchased other kind of more normal VPS from xHosts with "only" 30 TB monthly traffic was so good that it made its way into my top-3 lists in my "who's the bestest?" shootout thread.
Based on a bit over 30 runs btw.
But some inner voice told me to be a bit mistrusting because not only this VPS offers almost five times(!) the traffic for the same annual price but it also comes with a Ryzen. So logic suggested to look out for a weakness ...
And a weakness will be found, read on ...
First sysinfo, processor, and memory.
Hmm, very nice indeed and in fact even significantly faster than the other (Epyc based) VPS. Also, all the desired flags are there although I don't know why one one would want to put a nested VPS within a single-core VPS.
But whatever, not that even more performance (than the other VPS) was needed, but it's there and that's very nice. Well done xHosts!
Now let's look at the disk. Is it decent as well?
Yep, it does. OK, it's a bee stick less performant but it still is in the >= 20 MB/s and >= 5000 IOps class which I consider very decent for a really cheap VPS.
So: passed with colours!
Well, by now it's clear, where this promo VPS's weakness is: in the connectivity. But how grave is it? Let's have a look.
As usual, split into continents/regions. And keep in mind that what you see is the best one where there are multiple targets.
Europe - Yuck! That's a disappointing result set compared to the other xHosts promo VPS. Yes, there are some decent results and e.g. all 3 CH targets worked OK although one was very significantly weaker, but overall these numbers are mediocre for xHosts.
Asia/Oceania - Similar story, one brutal example being webwerks India, which achieved an even purer result than garuda and both are utterly crappy. On a positive and somewhat weird note both SGP and HKG again achieved decent results and Ozzyland results are kind of OK too (and all targets reached).
Africa - Nairobi actually quite good but South Africa in between mediocre and meeh
America - Partially a nice surprise with quite a few targets achieving significantly over 100 Mb/s. On the other hand tough the NYC and Ashburn targets, incl. @Clouvider - who while not being THE best actually is really well connected - and quite a few other targets show measly results. Rackspace and Leaseweb are less crappy but largely not great either.
WHY? Why when xHosts actually is well connected as the other, the "normal" VPS clearly - and winningly! - shows?
So, I took a deeper look and compared the routing of both VPS. Nope, that's not the issue, both basically use the same routes, give or take a hop or a few milliseconds. Yet, e.g. Clouvider NYC had over 10 times better results a few weeks ago when I benchmarked the "normal" xHosts VPS.
So I guess two potential culprits remain, either - for whatever reason - drastically worse throughput in general, i.e. not xHosts' fault but some carriers' -or- xHosts is limiting throughput.
Easy to find out. I just ran a fresh, albeit only single, run on the normal VPS. The result? Like known from (the other, the "normal") xHosts VPS. High to very high throughput within Europe, and NYC, as one example for NA, also is about 150 Mb/s.
So it seems the not at all announced price to pay for the extremely high traffic volume is that throughput is limited, and quite severely so I'd say.
TL;DR: It was clear that somewhere corners had to be cut, no surprise there. And I can understand it and wouldn't have a problem with it if the limiting were more reasonable, like "not 150 Mb/s to NYC but only 50 Mb/s (about 2/3 less) or 75 Mb/s (about 50% less). The way xHosts did it leaves us/me with a VPS that looked great but whose usability actually is quite questionable.
So, not recommended, sorry. I strongly suggest xHosts find a somewhat better compromise. Why for example Ryzen, why not an E5v4 but with unlimited or only mildly limited throughput? The way it is this box clearly isn't a winner nor a good deal (and I apologize for lauding the deal in xHosts' promo thread).