All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
HostHatch 9 locations Mega-benchmark/review
First, a big THANK YOU to HostHatch who not only provided 11 VPS in 9 locations to me but, even better, offered to play it in a different and arguably more trustworthy way: they gave me enough credits to "buy" all the VPSs I wanted to benchmark and test with the - very welcome and positive - intention that I should and could be sure to not deal with somehow "prepared" systems but to get my test VPSs just like any customer would get theirs. They told me that they'd like to see all of their (current) locations benchmarked but that was in no way binding; I was absolutely free to choose whatever I liked within the (very generous) budget provided. Kudos for that @hosthatch!
What I chose was a dual dedicated core, 8 GB mem., 60 GB NVMe. 4 TB traffic VPS in each of the 9 locations, plus their smallest VPS (one "50% dedicated" (whatever that means) core, 2 GB mem., 20 GB NVMe, 1 TB traffic) in NL, plus a 250 GB storage VPS (1 (smaller) core, 512 MB mem., 250 GB HDD, 1 TB traffic) in Zürich, so I tested 11 VPS in total. All of the KVM based btw.
To avoid trouble with Vanilla and/or CloudF#%&! I'll split the reviews/benchmarks into multiple posts with 2 locations each, first the american ones, then the european ones, both in alphabetical order, except for the first one which I'll put into this "master post", followed finally by the small and the storage VPS.
Btw, following some recent (justified) criticism/suggestion by @Falzo now all tests have units in brackets with them, e.g. "[MB/s:]". Let me also repeat that the 'webping' tests do not (so much) test the reviewed system but rather the test target (where a 100 MB download is attempted from); I came to include this in v.2 of my benchmark program after I had noticed that no matter how carefully one selects test targets there's always some random involved, e.g. systems which always responded quickly suddenly react like a snail and possibly are back to normal after a while. So, when you see either download speed or ping times keep that in mind and look at the web ping too. If it's considerably higher than the ping chances are that the target is PMSing and the reviewed VPS is innocent.
First, a general remark: HostHatch seem to have their own panel and that panel works fine and it's easy to get used to it except for one potential trap (into which I walked at the beginning): when ordering/selecting a VPS after clicking on the distribution there are two additional quite small and greyish fields to click or select some value from.So, look closely and attentatively when ordering.
About the only criticism I have with HostHatch and their panel is a double one related to the console (in the panel): (a) please offer a way to chose from at least a couple of languages for the console keyboard. It's immensely unnerving to almost blindly (that's how it feels) type on an unknown keyboard, and (b) please provide a way to "undock" the console window -and- to be able to click on another panel section, e.g. to choose another ISO image while the console is open.
Being at ISOs: compliments, HostHatch, for enabling us to upload and then use (install from, ...) any ISO we please. That's very nice and worth mentioning because many providers don't offer that or it's not properly working and one must open a ticket. Not so with Hosthatch, that's great!
Btw, each and every VPS was benchmarked at least about 120 times during about a week of testing, at different times of day and at unpredictable times (controlled by a (good quality) random driven benchmark driver). The only thing known was that a VPS wouldn't be tested more than once per half hour and not less frequent than every 90 minutes, all around the clock. So, if there are busy hours and/or weak usage hours one would see that (and trust me, I've seen really wild jumps depending on the hour of day with some providers).
Final "intro" remark: HostHatch's support was friendly, helpful, and responding quickly (typ. between some minutes and half an hour reaction time) and what they answered wasn't worthless canned BS but concrete and directly related to the question. If and when the solution was them needing to do something, the response was "problem solved" and the necessary had already been done. In short: very good support (but as always with a grain of salt because they of course knew who I was and that I was reviewing them).
So, here is the benchmark data of the first american location, Chicago:
Machine: amd64, Arch.: amd64, Model: Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz
OS, version: FreeBSD 12.2, Mem.: 7.986 GB
CPU - Cores: 2, Family/Model/Stepping: 6/62/4
Cache: 32K/32K L1d/L1i, 2M L2, 16M 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 ss sse3 pclmulqdq ssse3 cx16 pcid
sse4_1 sse4_2 x2apic popcnt tsc_deadline **aes** xsave osxsave avx f16c
rdrnd hypervisor
Ext. Flags: fsgsbase tsc_adjust smep erms syscall nx pdpe1gb rdtscp lm lahf_lm
ProcMem SC [MB/s]: avg 304.2 - min 298.5 (98.1 %), max 309.7 (101.8 %)
ProcMem MC [MB/s]: avg 616.9 - min 593.4 (96.2 %), max 637.1 (103.3 %)
--- Disk - Buffered ---
Write seq. [MB/s]: avg 1345.33 - min 950.62 (70.7%), max 1547.67 (115.0%)
Write rnd. [MB/s]: avg 4764.86 - min 3116.11 (65.4%), max 6062.18 (127.2%)
Read seq. [MB/s]: avg 4941.64 - min 3254.84 (65.9%), max 5825.25 (117.9%)
Read rnd. [MB/s]: avg 6799.30 - min 4987.47 (73.4%), max 8506.75 (125.1%)
--- Disk - Sync/Direct ---
Write seq. [MB/s]: avg 94.18 - min 60.83 (64.6%), max 104.51 (111.0%)
Write rnd. [MB/s]: avg 186.28 - min 118.11 (63.4%), max 209.69 (112.6%)
Read seq. [MB/s]: avg 3261.24 - min 2495.94 (76.5%), max 3685.33 (113.0%)
Read rnd. [MB/s]: avg 383.10 - min 198.93 (51.9%), max 415.66 (108.5%)
--- Network ---
US LAX lax.download.datapacket.com
DL [Mb/s]: avg 113.18 - min 102.32 (90.4%), max 119.43 (105.5%)
Ping [ms]: avg 57.0 - min 55.3 (96.9%), max 58.3 (102.2%)
Web ping [ms]: avg 58.3 - min 55.6 (95.4%), max 126.6 (217.3%)
NO OSL speedtest.osl01.softlayer.com
DL [Mb/s]: avg 54.63 - min 50.39 (92.2%), max 57.73 (105.7%)
Ping [ms]: avg 105.3 - min 105.2 (99.9%), max 106.2 (100.9%)
Web ping [ms]: avg 151.7 - min 105.2 (69.4%), max 1447.7 (954.5%)
US SJC speedtest.sjc01.softlayer.com
DL [Mb/s]: avg 116.59 - min 65.61 (56.3%), max 135.14 (115.9%)
Ping [ms]: avg 47.7 - min 47.6 (99.8%), max 48.5 (101.7%)
Web ping [ms]: avg 94.1 - min 47.6 (50.6%), max 2081.2 (2211.0%)
JP TOK speedtest.tokyo2.linode.com
DL [Mb/s]: avg 46.13 - min 37.07 (80.4%), max 49.18 (106.6%)
Ping [ms]: avg 135.3 - min 134.9 (99.7%), max 136.0 (100.6%)
Web ping [ms]: avg 135.4 - min 134.9 (99.6%), max 137.1 (101.3%)
IT MIL speedtest.mil01.softlayer.com
DL [Mb/s]: avg 52.93 - min 47.21 (89.2%), max 56.75 (107.2%)
Ping [ms]: avg 107.0 - min 106.2 (99.3%), max 108.7 (101.6%)
Web ping [ms]: avg 129.0 - min 106.4 (82.5%), max 1455.1 (1128.2%)
FR PAR speedtest.par01.softlayer.com
DL [Mb/s]: avg 65.06 - min 63.17 (97.1%), max 67.35 (103.5%)
Ping [ms]: avg 93.6 - min 92.7 (99.0%), max 94.6 (101.0%)
Web ping [ms]: avg 109.4 - min 92.7 (84.8%), max 778.1 (711.5%)
SG SGP mirror.sg.leaseweb.net
DL [Mb/s]: avg 30.17 - min 4.76 (15.8%), max 33.85 (112.2%)
Ping [ms]: avg 194.1 - min 189.1 (97.4%), max 201.3 (103.7%)
Web ping [ms]: avg 194.2 - min 189.1 (97.4%), max 201.3 (103.7%)
BR SAO speedtest.sao01.softlayer.com
DL [Mb/s]: avg 35.46 - min 2.44 (6.9%), max 47.24 (133.2%)
Ping [ms]: avg 130.0 - min 129.6 (99.7%), max 131.8 (101.4%)
Web ping [ms]: avg 194.7 - min 129.7 (66.6%), max 2522.8 (1295.5%)
IN CHN speedtest.che01.softlayer.com
DL [Mb/s]: avg 23.40 - min 13.04 (55.7%), max 26.36 (112.6%)
Ping [ms]: avg 27185.8 - min 229.8 (0.8%), max 1349557.1 (4964.2%)
Web ping [ms]: avg 27223.4 - min 231.5 (0.9%), max 1349557.1 (4957.3%)
US WDC mirror.wdc1.us.leaseweb.net
DL [Mb/s]: avg 248.19 - min 102.19 (41.2%), max 269.32 (108.5%)
Ping [ms]: avg 21.4 - min 21.3 (99.4%), max 22.3 (104.1%)
Web ping [ms]: avg 21.5 - min 21.3 (99.2%), max 22.3 (103.8%)
DE FRA speedtest.fra02.softlayer.com
DL [Mb/s]: avg 55.64 - min 48.09 (86.4%), max 60.61 (108.9%)
Ping [ms]: avg 98.2 - min 97.9 (99.7%), max 99.7 (101.5%)
Web ping [ms]: avg 106.5 - min 97.9 (92.0%), max 988.4 (928.4%)
RU MOS speedtest.hostkey.ru
DL [Mb/s]: avg 41.06 - min 0.35 (0.9%), max 47.66 (116.1%)
Ping [ms]: avg 138.6 - min 133.5 (96.3%), max 146.1 (105.4%)
Web ping [ms]: avg 140.5 - min 133.6 (95.1%), max 159.7 (113.7%)
US DAL speedtest.dal05.softlayer.com
DL [Mb/s]: avg 236.03 - min 207.09 (87.7%), max 265.37 (112.4%)
Ping [ms]: avg 24.0 - min 23.0 (95.7%), max 25.6 (106.6%)
Web ping [ms]: avg 26.2 - min 23.0 (87.8%), max 293.1 (1118.8%)
UK LON speedtest.lon02.softlayer.com
DL [Mb/s]: avg 68.82 - min 0.00 (0.0%), max 73.36 (106.6%) - (http error: -10)
Ping [ms]: avg 89.3 - min 87.1 (97.5%), max 90.4 (101.2%)
Web ping [ms]: avg 194.6 - min 87.2 (44.8%), max 2219.3 (1140.7%)
US NYC nyc.download.datapacket.com
DL [Mb/s]: avg 292.65 - min 176.53 (60.3%), max 358.53 (122.5%)
Ping [ms]: avg 19.6 - min 17.2 (87.9%), max 20.9 (106.8%)
Web ping [ms]: avg 20.4 - min 17.3 (84.6%), max 64.3 (314.5%)
RO BUC 185.183.99.8
DL [Mb/s]: avg 46.93 - min 43.67 (93.1%), max 52.51 (111.9%)
Ping [ms]: avg 135.6 - min 124.4 (91.8%), max 137.6 (101.5%)
Web ping [ms]: avg 141.2 - min 124.4 (88.1%), max 532.3 (376.9%)
CN_HK mirror.hk.leaseweb.net
DL [Mb/s]: avg 35.43 - min 18.72 (52.8%), max 36.81 (103.9%)
Ping [ms]: avg 175.5 - min 175.4 (100.0%), max 176.4 (100.5%)
Web ping [ms]: avg 175.6 - min 175.4 (99.9%), max 178.3 (101.5%)
So, while the processor may not sound impressiv, it actually is. For one about 300 is less than a Ryzen 3000 but not really by that much (the Ryzen 3950X processor in a VPS I recently tested performed avg. ca. 370 that is, just about 20% better). Ca. 300 is a very decent result and both benchmarks (single and multi core) showed very high consistency with less than 5% spread in both directions. Plus, those 2 vCores really are dedicated. Very positive impression there.
The disk are probably the fastest I've seen so far in any VPS, and I've seen (and benchmarked) quite a few disks of all sorts. What really impressed me about them was their almost stubborn consistency. With only one exception, and still a respectable one, the negative spread always stayed with about one third. And btw, when I say that the exception is still respectable I mean it; it's around 50% where other disks I benchmarked, and not bad ones mind you, occasionally drop down about 75%. Also please note that when I write "min" (in the result data) that is the minimum of all benchmark runs over about a week (well over 120 runs). Finally, also note that the avg. speed always is much closer to max than to min (or even to the mid value). Very nice disks indeed!
As for the network, well it's not bad, really not, but neither is it something to write home or brag about. But here the utility of the web ping becomes visible: Wherever you see a mediocre or even poor result the web ping is way over 100% too, which means that the other side (the server we tried to download from) was sleeping. Does that mean that all is good and great? Certainly not; while bad luck plays a role there is no denying that the DC upstream plays the major role.
Frankly, if HostHatch asked me for advice it would be something along the lines of "Hey, (a) you are HostHatch and you have sunk a ton of money into really, really nice hardware, (b) forget about everyone and his dog talking about Ryzen. Your hardware is great, just continue on that road, but, damn go get a better carrier!"
Also: HostHatch isn't playing in the "we are the cheapest!!!" game. Their game is "You get great hardware, consistency, unearthly fast disks, good support, and all of that for a very reasonable price!". AFAIC a provider like that is drawn down by a mediocre carrier like M247. Pardon my being frank.
So, yes, others sell you a 1 vCore cheap, cheap VPS for half of what HostHatch asks for - but: that super cheapo provider will run your VPS on a processor with half the performance, maybe no AES, 512 MB mem (instead of 2 GB), a 5 to 10 GB SSD (if you are lucky), support usually not worth mentioning, ... while HostHatch's smallest VPS comes very well equipped and with good support.
Comments
.
part 2, first Los Angeles and then New York -
Machine: amd64, Arch.: amd64, Model: Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz
OS, version: FreeBSD 12.2, Mem.: 7.986 GB
CPU - Cores: 2, Family/Model/Stepping: 6/62/4
Cache: 32K/32K L1d/L1i, 2M L2, 16M 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 ss sse3 pclmulqdq ssse3 cx16 pcid
sse4_1 sse4_2 x2apic popcnt tsc_deadline aes xsave osxsave avx f16c
rdrnd hypervisor
Ext. Flags: fsgsbase tsc_adjust smep erms syscall nx pdpe1gb rdtscp lm lahf_lm
ProcMem SC [MB/s]: avg 300.8 - min 292.6 (97.3 %), max 305.3 (101.5 %)
ProcMem MC [MB/s]: avg 611.7 - min 579.0 (94.7 %), max 633.5 (103.6 %)
--- Disk - Buffered ---
Write seq. [MB/s]: avg 1228.08 - min 1127.45 (91.8%), max 1353.44 (110.2%)
Write rnd. [MB/s]: avg 4185.88 - min 3408.54 (81.4%), max 5069.22 (121.1%)
Read seq. [MB/s]: avg 4191.44 - min 3554.51 (84.8%), max 4833.94 (115.3%)
Read rnd. [MB/s]: avg 5710.81 - min 3831.42 (67.1%), max 6866.29 (120.2%)
--- Disk - Sync/Direct ---
Write seq. [MB/s]: avg 75.84 - min 49.34 (65.1%), max 91.51 (120.7%)
Write rnd. [MB/s]: avg 145.95 - min 96.26 (66.0%), max 184.45 (126.4%)
Read seq. [MB/s]: avg 2795.32 - min 2178.40 (77.9%), max 3211.25 (114.9%)
Read rnd. [MB/s]: avg 286.38 - min 212.88 (74.3%), max 370.49 (129.4%)
--- Network ---
US LAX lax.download.datapacket.com
DL [Mb/s]: avg 1093.02 - min 1020.52 (93.4%), max 1122.26 (102.7%)
Ping [ms]: avg 0.6 - min 0.5 (81.7%), max 1.0 (163.4%)
Web ping [ms]: avg 1.0 - min 0.5 (49.8%), max 29.4 (2929.7%)
NO OSL speedtest.osl01.softlayer.com
DL [Mb/s]: avg 38.49 - min 28.42 (73.8%), max 40.84 (106.1%)
Ping [ms]: avg 157.1 - min 154.9 (98.6%), max 213.4 (135.8%)
Web ping [ms]: avg 178.7 - min 155.0 (86.7%), max 1498.8 (838.6%)
US SJC speedtest.sjc01.softlayer.com
DL [Mb/s]: avg 465.79 - min 268.00 (57.5%), max 662.47 (142.2%)
Ping [ms]: avg 8.1 - min 8.0 (98.6%), max 9.9 (122.0%)
Web ping [ms]: avg 143.7 - min 8.1 (5.6%), max 2578.3 (1794.7%)
JP TOK speedtest.tokyo2.linode.com
DL [Mb/s]: avg 53.84 - min 37.14 (69.0%), max 60.88 (113.1%)
Ping [ms]: avg 120.7 - min 110.6 (91.6%), max 130.0 (107.7%)
Web ping [ms]: avg 121.3 - min 110.8 (91.3%), max 130.0 (107.1%)
IT MIL speedtest.mil01.softlayer.com
DL [Mb/s]: avg 37.95 - min 35.73 (94.1%), max 40.51 (106.7%)
Ping [ms]: avg 153.8 - min 151.9 (98.7%), max 155.3 (101.0%)
Web ping [ms]: avg 192.3 - min 152.0 (79.0%), max 1708.2 (888.2%)
FR PAR speedtest.par01.softlayer.com
DL [Mb/s]: avg 43.11 - min 29.37 (68.1%), max 46.19 (107.1%)
Ping [ms]: avg 142.7 - min 138.1 (96.8%), max 145.6 (102.0%)
Web ping [ms]: avg 145.8 - min 139.6 (95.8%), max 258.5 (177.3%)
SG SGP mirror.sg.leaseweb.net
DL [Mb/s]: avg 35.02 - min 26.57 (75.9%), max 38.46 (109.8%)
Ping [ms]: avg 178.8 - min 161.5 (90.3%), max 223.4 (124.9%)
Web ping [ms]: avg 179.7 - min 161.6 (90.0%), max 223.4 (124.4%)
BR SAO speedtest.sao01.softlayer.com
DL [Mb/s]: avg 27.05 - min 2.22 (8.2%), max 36.50 (134.9%)
Ping [ms]: avg 172.5 - min 169.4 (98.2%), max 174.7 (101.3%)
Web ping [ms]: avg 248.3 - min 170.8 (68.8%), max 2153.5 (867.3%)
IN CHN speedtest.che01.softlayer.com
DL [Mb/s]: avg 24.04 - min 13.77 (57.3%), max 28.92 (120.3%)
Ping [ms]: avg 15133.7 - min 209.1 (1.4%), max 1337027.0 (8834.8%)
Web ping [ms]: avg 15163.1 - min 210.4 (1.4%), max 1337027.0 (8817.6%)
US WDC mirror.wdc1.us.leaseweb.net
DL [Mb/s]: avg 98.40 - min 47.07 (47.8%), max 109.82 (111.6%)
Ping [ms]: avg 62.2 - min 62.1 (99.8%), max 62.5 (100.4%)
Web ping [ms]: avg 62.4 - min 62.1 (99.5%), max 64.4 (103.2%)
DE FRA speedtest.fra02.softlayer.com
DL [Mb/s]: avg 39.28 - min 33.58 (85.5%), max 41.71 (106.2%)
Ping [ms]: avg 147.2 - min 143.2 (97.3%), max 150.4 (102.2%)
Web ping [ms]: avg 202.3 - min 143.3 (70.8%), max 2285.7 (1129.8%)
RU MOS speedtest.hostkey.ru
DL [Mb/s]: avg 33.62 - min 11.25 (33.5%), max 37.02 (110.1%)
Ping [ms]: avg 181.8 - min 181.2 (99.7%), max 205.7 (113.1%)
Web ping [ms]: avg 183.1 - min 181.3 (99.0%), max 210.2 (114.8%)
US DAL speedtest.dal05.softlayer.com
DL [Mb/s]: avg 168.35 - min 111.88 (66.5%), max 206.88 (122.9%)
Ping [ms]: avg 29.5 - min 29.3 (99.5%), max 30.3 (102.8%)
Web ping [ms]: avg 37.1 - min 29.4 (79.2%), max 892.2 (2404.7%)
UK LON speedtest.lon02.softlayer.com
DL [Mb/s]: avg 44.92 - min 32.59 (72.5%), max 47.83 (106.5%)
Ping [ms]: avg 141.2 - min 136.7 (96.8%), max 143.0 (101.3%)
Web ping [ms]: avg 179.6 - min 139.2 (77.5%), max 1412.4 (786.4%)
US NYC nyc.download.datapacket.com
DL [Mb/s]: avg 90.91 - min 44.65 (49.1%), max 104.99 (115.5%)
Ping [ms]: avg 67.2 - min 60.9 (90.7%), max 74.2 (110.5%)
Web ping [ms]: avg 69.6 - min 60.9 (87.5%), max 86.6 (124.4%)
RO BUC 185.183.99.8
DL [Mb/s]: avg 36.27 - min 25.31 (69.8%), max 38.54 (106.3%)
Ping [ms]: avg 172.7 - min 170.3 (98.6%), max 255.3 (147.9%)
Web ping [ms]: avg 180.3 - min 170.4 (94.5%), max 360.8 (200.1%)
CN_HK mirror.hk.leaseweb.net
DL [Mb/s]: avg 39.38 - min 37.17 (94.4%), max 41.08 (104.3%)
Ping [ms]: avg 155.7 - min 151.6 (97.4%), max 161.2 (103.6%)
Web ping [ms]: avg 158.5 - min 151.9 (95.8%), max 169.1 (106.7%)
First, notice the data for Tokyo, Hongkong, Singapore! Quite decent, at least from my european perspective (we go to East Asia the other way around).
Otherwise, as I already said, almost boring. Stubbornly consistent, very nice processor and memory speed, unearthly fast NVMe disks, etc.
I'm getting mistrusting. Is Abdullah having an affair with my wife or wha the hell did he build HostHatch in way to make me fall asleep? Again: Stubbornly consistent, very nice processor and memory speed, unearthly fast NVMe disks, etc.
Oh wait, I gotcha Abdullah! I noticed something out of the order: this nodes disks are even a tiny wee bit faster than the ones in LA. I'm more and more convince of a sinister conspiracy against me ...
part 3, Europe - Vienna and Zürich (yes, I ordered them by country code not city name)
Ha! A weakling! ... Just joking. It's just that this processor runs at "only" 2.8 GHz (rather than 3.0).
Otherwise: booooooringly stubbornly consistent performance. I'm beginning to miss benchmarking them $2/mo VPSs from some providers which held a surprise every other line. So, so at 6 am, snarky at noon, etc ... Meeh
Side note: The "low" 2.8 GHz are easy to forgive considering that Vienna isn't exactly a major european internet hub. Frankly, I wondered why at all; both Bratislava and Milano would have been more attractive. Well, maybe the HostHatch guys love Sacher cake and wanted an excuse for occasional Vienna visits, hmmm ...
Now on to Zürich, CH
o0O(just checked in the bed room. Wife is sleeping. Alone.)
I've had enough of those boring benchmarks. I'll strike back! Take that, Abdullah -> Stubbornly consistent, very nice processor and memory speed, unearthly fast NVMe disks, etc.
part ? (sorry I was so bored that I fell asleep. Can someone tell me the number, please?!) Anyway this is NL and then Norway (yummy, yummy)
Machine: amd64, Arch.: amd64, Model: Intel(R) Xeon(R) CPU E5-2690 0 @ 2.90GHz
OS, version: FreeBSD 12.2, Mem.: 7.986 GB
CPU - Cores: 2, Family/Model/Stepping: 6/45/7
Cache: 32K/32K L1d/L1i, 256K L2, 20M 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 ss sse3 pclmulqdq ssse3 cx16 pcid
sse4_1 sse4_2 x2apic popcnt tsc_deadline ++aes++ xsave osxsave avx
hypervisor
Ext. Flags: tsc_adjust syscall nx pdpe1gb rdtscp lm lahf_lm
ProcMem SC [MB/s]: avg 247.7 - min 233.0 (94.1 %), max 255.9 (103.3 %)
ProcMem MC [MB/s]: avg 504.7 - min 484.6 (96.0 %), max 524.8 (104.0 %)
--- Disk - Buffered ---
Write seq. [MB/s]: avg 1332.09 - min 1122.53 (84.3%), max 1470.29 (110.4%)
Write rnd. [MB/s]: avg 4409.52 - min 3424.90 (77.7%), max 5292.69 (120.0%)
Read seq. [MB/s]: avg 4591.72 - min 3016.85 (65.7%), max 5677.71 (123.7%)
Read rnd. [MB/s]: avg 5917.05 - min 3768.05 (63.7%), max 7920.12 (133.9%)
--- Disk - Sync/Direct ---
Write seq. [MB/s]: avg 97.38 - min 65.55 (67.3%), max 104.51 (107.3%)
Write rnd. [MB/s]: avg 193.18 - min 121.87 (63.1%), max 207.98 (107.7%)
Read seq. [MB/s]: avg 2885.78 - min 1757.93 (60.9%), max 3586.50 (124.3%)
Read rnd. [MB/s]: avg 386.35 - min 227.63 (58.9%), max 419.43 (108.6%)
--- Network ---
US LAX lax.download.datapacket.com
DL [Mb/s]: avg 72.09 - min 39.66 (55.0%), max 281.33 (390.3%)
Ping [ms]: avg 128.8 - min 20.2 (15.7%), max 153.6 (119.2%)
Web ping [ms]: avg 131.1 - min 20.2 (15.4%), max 163.7 (124.9%)
NO OSL speedtest.osl01.softlayer.com
DL [Mb/s]: avg 196.29 - min 188.98 (96.3%), max 207.00 (105.5%)
Ping [ms]: avg 24.6 - min 24.2 (98.4%), max 25.4 (103.3%)
Web ping [ms]: avg 43.8 - min 24.2 (55.3%), max 1223.0 (2795.1%)
US SJC speedtest.sjc01.softlayer.com
DL [Mb/s]: avg 42.55 - min 40.51 (95.2%), max 44.38 (104.3%)
Ping [ms]: avg 149.9 - min 149.7 (99.9%), max 150.6 (100.5%)
Web ping [ms]: avg 239.3 - min 149.7 (62.5%), max 2062.6 (861.8%)
JP TOK speedtest.tokyo2.linode.com
DL [Mb/s]: avg 27.15 - min 23.31 (85.9%), max 28.95 (106.6%)
Ping [ms]: avg 238.4 - min 233.5 (98.0%), max 246.3 (103.3%)
Web ping [ms]: avg 239.6 - min 233.6 (97.5%), max 249.7 (104.2%)
IT MIL speedtest.mil01.softlayer.com
DL [Mb/s]: avg 221.07 - min 213.59 (96.6%), max 228.07 (103.2%)
Ping [ms]: avg 22.1 - min 22.0 (99.4%), max 22.7 (102.6%)
Web ping [ms]: avg 38.7 - min 22.0 (56.9%), max 1696.7 (4387.8%)
FR PAR speedtest.par01.softlayer.com
DL [Mb/s]: avg 426.28 - min 242.82 (57.0%), max 491.41 (115.3%)
Ping [ms]: avg 11.6 - min 11.4 (98.2%), max 12.1 (104.2%)
Web ping [ms]: avg 26.1 - min 11.4 (43.7%), max 607.7 (2331.4%)
SG SGP mirror.sg.leaseweb.net
DL [Mb/s]: avg 31.51 - min 20.71 (65.7%), max 40.75 (129.3%)
Ping [ms]: avg 208.7 - min 192.6 (92.3%), max 231.9 (111.1%)
Web ping [ms]: avg 235.5 - min 192.6 (81.8%), max 275.9 (117.2%)
BR SAO speedtest.sao01.softlayer.com
DL [Mb/s]: avg 30.97 - min 28.56 (92.2%), max 33.14 (107.0%)
Ping [ms]: avg 187.0 - min 184.0 (98.4%), max 189.8 (101.5%)
Web ping [ms]: avg 260.5 - min 184.1 (70.7%), max 2632.7 (1010.5%)
IN CHN speedtest.che01.softlayer.com
DL [Mb/s]: avg 22.39 - min 19.14 (85.5%), max 24.89 (111.1%)
Ping [ms]: avg 20715.0 - min 241.6 (1.2%), max 1250958.5 (6038.9%)
Web ping [ms]: avg 20749.9 - min 245.5 (1.2%), max 1250958.5 (6028.7%)
US WDC mirror.wdc1.us.leaseweb.net
DL [Mb/s]: avg 66.72 - min 39.97 (59.9%), max 74.09 (111.0%)
Ping [ms]: avg 88.8 - min 88.1 (99.2%), max 95.5 (107.6%)
Web ping [ms]: avg 88.9 - min 88.1 (99.1%), max 96.0 (107.9%)
DE FRA speedtest.fra02.softlayer.com
DL [Mb/s]: avg 613.38 - min 425.28 (69.3%), max 748.96 (122.1%)
Ping [ms]: avg 7.0 - min 6.8 (97.3%), max 7.4 (105.8%)
Web ping [ms]: avg 10.1 - min 6.8 (67.5%), max 343.3 (3409.9%)
RU MOS speedtest.hostkey.ru
DL [Mb/s]: avg 105.40 - min 1.09 (1.0%), max 120.54 (114.4%)
Ping [ms]: avg 50.9 - min 17.6 (34.6%), max 55.7 (109.4%)
Web ping [ms]: avg 53.4 - min 49.4 (92.5%), max 74.4 (139.3%)
US DAL speedtest.dal05.softlayer.com
DL [Mb/s]: avg 53.20 - min 51.15 (96.1%), max 54.99 (103.4%)
Ping [ms]: avg 114.5 - min 114.3 (99.8%), max 115.3 (100.7%)
Web ping [ms]: avg 115.7 - min 114.3 (98.8%), max 187.9 (162.4%)
UK LON speedtest.lon02.softlayer.com
DL [Mb/s]: avg 515.72 - min 0.00 (0.0%), max 695.72 (134.9%) - (http error: -10)
Ping [ms]: avg 8.3 - min 8.2 (98.3%), max 8.6 (103.1%)
Web ping [ms]: avg 53.3 - min 8.2 (15.4%), max 2061.2 (3869.0%)
US NYC nyc.download.datapacket.com
DL [Mb/s]: avg 103.36 - min 42.55 (41.2%), max 276.64 (267.6%)
Ping [ms]: avg 69.7 - min 20.2 (29.0%), max 80.8 (115.9%)
Web ping [ms]: avg 71.1 - min 20.2 (28.4%), max 82.0 (115.3%)
RO BUC 185.183.99.8
DL [Mb/s]: avg 159.31 - min 110.49 (69.4%), max 180.75 (113.5%)
Ping [ms]: avg 32.7 - min 32.4 (99.1%), max 34.9 (106.8%)
Web ping [ms]: avg 37.6 - min 32.5 (86.5%), max 268.0 (713.4%)
CN_HK mirror.hk.leaseweb.net
DL [Mb/s]: avg 33.76 - min 31.88 (94.4%), max 36.02 (106.7%)
Ping [ms]: avg 182.3 - min 176.2 (96.7%), max 190.0 (104.2%)
Web ping [ms]: avg 186.6 - min 176.5 (94.6%), max 194.2 (104.1%)
Stubbornly consistent, very nice processor and memory speed, unearthly fast NVMe disks, etc.
I'm thinking about establishing a benchmarkers union ...
Now, the Norway VPS
Stubbornly consistent, very nice processor and memory speed, unearthly fast NVMe disks, etc.
I started to mark the AES flag with double plus signs rather than double asterisks as an act of revenge against them HostHatch guys creating 9 locations (!!) with consistently good performance just to make my reviewer life miserable!
Machine: amd64, Arch.: amd64, Model: Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
OS, version: FreeBSD 12.2, Mem.: 7.986 GB
CPU - Cores: 2, Family/Model/Stepping: 6/62/4
Cache: 32K/32K L1d/L1i, 2M L2, 16M 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 ss sse3 pclmulqdq ssse3 cx16 pcid
sse4_1 sse4_2 x2apic popcnt tsc_deadline aes xsave osxsave avx f16c
rdrnd hypervisor
Ext. Flags: fsgsbase tsc_adjust smep erms syscall nx pdpe1gb rdtscp lm lahf_lm
ProcMem SC [MB/s]: avg 264.1 - min 161.4 (61.1 %), max 280.1 (106.1 %)
ProcMem MC [MB/s]: avg 531.6 - min 374.1 (70.4 %), max 572.1 (107.6 %)
--- Disk - Buffered ---
Write seq. [MB/s]: avg 988.50 - min 305.02 (30.9%), max 1383.26 (139.9%)
Write rnd. [MB/s]: avg 4529.74 - min 2269.07 (50.1%), max 5835.27 (128.8%)
Read seq. [MB/s]: avg 4174.60 - min 2068.99 (49.6%), max 5411.82 (129.6%)
Read rnd. [MB/s]: avg 6571.20 - min 2937.59 (44.7%), max 8038.91 (122.3%)
--- Disk - Sync/Direct ---
Write seq. [MB/s]: avg 56.96 - min 12.05 (21.2%), max 84.20 (147.8%)
Write rnd. [MB/s]: avg 117.37 - min 29.80 (25.4%), max 182.09 (155.1%)
Read seq. [MB/s]: avg 2667.38 - min 1508.87 (56.6%), max 3153.67 (118.2%)
Read rnd. [MB/s]: avg 238.72 - min 93.36 (39.1%), max 314.91 (131.9%)
--- Network ---
US LAX lax.download.datapacket.com
DL [Mb/s]: avg 57.13 - min 0.85 (1.5%), max 179.26 (313.8%)
Ping [ms]: avg 10349.0 - min 35.6 (0.3%), max 1096963.1 (10599.7%)
Web ping [ms]: avg 10350.7 - min 35.6 (0.3%), max 1096963.1 (10597.9%)
NO OSL speedtest.osl01.softlayer.com
DL [Mb/s]: avg 550.29 - min 234.95 (42.7%), max 664.40 (120.7%)
Ping [ms]: avg 8.0 - min 7.8 (97.2%), max 18.6 (231.8%)
Web ping [ms]: avg 38.2 - min 7.9 (20.7%), max 1356.9 (3550.7%)
US SJC speedtest.sjc01.softlayer.com
DL [Mb/s]: avg 38.50 - min 32.57 (84.6%), max 40.60 (105.5%)
Ping [ms]: avg 156.9 - min 156.8 (99.9%), max 157.5 (100.4%)
Web ping [ms]: avg 227.0 - min 157.1 (69.2%), max 2494.4 (1098.9%)
JP TOK speedtest.tokyo2.linode.com
DL [Mb/s]: avg 23.34 - min 19.77 (84.7%), max 25.64 (109.8%)
Ping [ms]: avg 4858.6 - min 270.0 (5.6%), max 546229.6 (11242.5%)
Web ping [ms]: avg 4860.8 - min 270.0 (5.6%), max 546229.6 (11237.5%)
IT MIL speedtest.mil01.softlayer.com
DL [Mb/s]: avg 158.61 - min 110.33 (69.6%), max 170.01 (107.2%)
Ping [ms]: avg 29.9 - min 29.7 (99.5%), max 31.3 (104.8%)
Web ping [ms]: avg 56.3 - min 29.8 (53.0%), max 2203.4 (3916.9%)
FR PAR speedtest.par01.softlayer.com
DL [Mb/s]: avg 168.79 - min 57.36 (34.0%), max 195.56 (115.9%)
Ping [ms]: avg 30.1 - min 29.9 (99.4%), max 30.6 (101.8%)
Web ping [ms]: avg 38.4 - min 29.9 (77.9%), max 715.2 (1864.1%)
SG SGP mirror.sg.leaseweb.net
DL [Mb/s]: avg 18.22 - min 13.22 (72.6%), max 20.07 (110.2%)
Ping [ms]: avg 353.5 - min 351.1 (99.3%), max 398.9 (112.8%)
Web ping [ms]: avg 355.3 - min 351.1 (98.8%), max 398.9 (112.3%)
BR SAO speedtest.sao01.softlayer.com
DL [Mb/s]: avg 29.62 - min 25.42 (85.8%), max 30.96 (104.5%)
Ping [ms]: avg 203.6 - min 201.4 (98.9%), max 206.8 (101.6%)
Web ping [ms]: avg 272.6 - min 201.5 (73.9%), max 2482.4 (910.6%)
IN CHN speedtest.che01.softlayer.com
DL [Mb/s]: avg 29.06 - min 22.09 (76.0%), max 32.70 (112.5%)
Ping [ms]: avg 11319.2 - min 187.9 (1.7%), max 1335274.9 (11796.6%)
Web ping [ms]: avg 11369.0 - min 188.2 (1.7%), max 1335274.9 (11744.9%)
US WDC mirror.wdc1.us.leaseweb.net
DL [Mb/s]: avg 57.50 - min 0.29 (0.5%), max 62.84 (109.3%)
Ping [ms]: avg 103.3 - min 102.9 (99.6%), max 117.5 (113.7%)
Web ping [ms]: avg 104.1 - min 102.9 (98.9%), max 139.2 (133.8%)
DE FRA speedtest.fra02.softlayer.com
DL [Mb/s]: avg 229.21 - min 105.24 (45.9%), max 234.91 (102.5%)
Ping [ms]: avg 21.0 - min 20.9 (99.7%), max 22.1 (105.4%)
Web ping [ms]: avg 65.5 - min 20.9 (31.9%), max 1659.5 (2532.6%)
RU MOS speedtest.hostkey.ru
DL [Mb/s]: avg 106.93 - min 33.74 (31.6%), max 273.92 (256.2%)
Ping [ms]: avg 19.3 - min 18.8 (97.5%), max 38.4 (199.1%)
Web ping [ms]: avg 20.5 - min 19.0 (92.6%), max 57.3 (279.2%)
US DAL speedtest.dal05.softlayer.com
DL [Mb/s]: avg 45.43 - min 32.22 (70.9%), max 47.20 (103.9%)
Ping [ms]: avg 136.8 - min 136.6 (99.9%), max 137.1 (100.3%)
Web ping [ms]: avg 137.0 - min 136.6 (99.7%), max 153.4 (111.9%)
UK LON speedtest.lon02.softlayer.com
DL [Mb/s]: avg 207.93 - min 87.07 (41.9%), max 227.23 (109.3%)
Ping [ms]: avg 24.7 - min 24.5 (99.1%), max 26.2 (106.0%)
Web ping [ms]: avg 83.1 - min 24.5 (29.5%), max 2122.6 (2553.0%)
US NYC nyc.download.datapacket.com
DL [Mb/s]: avg 82.62 - min 47.41 (57.4%), max 180.08 (218.0%)
Ping [ms]: avg 5908.7 - min 35.5 (0.6%), max 693735.5 (11740.8%)
Web ping [ms]: avg 5915.3 - min 35.6 (0.6%), max 693735.5 (11727.8%)
RO BUC 185.183.99.8
DL [Mb/s]: avg 114.41 - min 51.78 (45.3%), max 129.64 (113.3%)
Ping [ms]: avg 49.9 - min 49.6 (99.4%), max 54.0 (108.2%)
Web ping [ms]: avg 52.0 - min 49.6 (95.4%), max 144.1 (277.1%)
CN_HK mirror.hk.leaseweb.net
DL [Mb/s]: avg 22.28 - min 0.30 (1.3%), max 25.26 (113.4%)
Ping [ms]: avg 291.8 - min 267.5 (91.7%), max 306.5 (105.0%)
Web ping [ms]: avg 292.6 - min 267.5 (91.4%), max 322.6 (110.3%)
OK, what am I supposed to do? How am I supposed to write a review on VPSs in 9 locations when they seems to somehow switch one and the same VPS all around the globe just to make my review look boring?
Ha! Not with me. I'll fight back! -> Oh, wait, almost 30 Mb/s even to Chennai, IN and Sao Paolo, BR? Damn, give me something to complain about!
Now, on to London
Stubbornly consistent, very nice processor and memory speed, unearthly fast NVMe disks, etc.
Revenge level increased to marking the AES flag with both a plus sign and an asterisk. That'll teach them a lesson.
Summary review:
HostHatch is both unwilling and incapable of offering some variety and adventure in their VPS hosting plans. ALL their VPS, no matter the location, are invariably very consistent, performant, and have unearthly fast NVMe disks. Plus, their support is good.
I'd like to say something bad about them, I really, really do, and I'd love to exact revenge on them for boring me mercilessly, but there is no hook, not even a small one to attach my revenge to. They have broken me, I'm a reviewer in tears. But there are still the two other VPS, maybe there is an opportunity for revenge there ... (dark laughter and crying at the same time)
Oh, how bitter sweet revenge tastes, oh how I love that. Finally my chance for revenge: I broke out of the continent.country pattern and put Zürich first ...
Wait, WHAT? Hello? Earth to HostHatch! this is a small storage VPS and I'm expecting something like single core performance in the 150 range or maybe 200 if you you want to brag - and what do you do? You put the same bloody processor and memory in that box and hence the same bloody good performance as in your normal VPSs!!! And to add insult to injury (sniff, cry) you don't put some measly Sata drives in it with some run of the mill Raid 5 or 6 controller but you evidently use very performant disks into that storage node along with a shockingly performant controller (and probably some serious NMVe caching too and offer way better disk performance on a measly small storage VPS than many providers offer with their top of the line VPS.
How on earth am I supposed to punish you for boring me, when you decide to suddenly enchant me?
Well, on to the small NL VPS
Finally at least a sign of some good will from HostHatch. This VPS's processor and memory is less performant than in their mid size and large VPSs. Plus the disk is just a decent SSD.
But hell, this is their cheapest, smallest VPS and it's still on par with quite a few providers mid-size VPS offers. Plus with 2 GB it has way too much memory for a cheapest, smallest box!
Verdict: The HostHatch guys are either unwilling or not capable to provide a VPS similar to most VPS out there from other providers. Even their smallest, cheapest VPS fails in that regard. I can only recommend HostHatch products to people who want very good good performance with very high performance NVMes and that at a very reasonable price. To make it worse their support is fast, competent and friendly and offers little surface for complaints.
Plus, and worst of all, the performance of their product is so consistently great that reviewers are bound to fall asleep.
P.S. Please note the expression of my dismay about their unwillingness to offer me some surface for complaints. I utterly failed and I'll leave now to cry into my pillow.
my bad, very long english
??
informative and comprehensive. good for me that is quite boring with YABS.. thanks @jsg and well done @hosthatch
Thanks ... and I sincerely hope that my attempt at humor didn't fail and the review is misunderstood.
What I meant to say, albeit slightly disguised in my attempt at being funny, is HOSTHATCH OFFERS GREAT PRODUCTS - and I mean really, really nice ones.
Very detailed and I believe they're within provider's expectation about his services. I wish data could be visualized in some nice graphs instead of a large dump of numbers. For example, one graph to show disk read/write of boxes across all locations to show both performance and consistency.
One question tho: The numbers are telling me that read/write to disk is faster than read/write to RAM, isn't it?
I love this thread. What a great review. Thanks for putting all this effort into it.
I am sorry, I really am but (a) trust me, running those benchmarks and compiling hundreds of result data sets is already lots of work, and (b) me doing graphics, even simple ones, would end in a desaster making your eyes bleed. I'm about as untalented and incapable with anything graphics as is possible.
Also think about those (not few) users who want specific data such as "I don't care about New York or London, what I need and look for is connectivity to (e.g.) India".
No, the "ProcMem" numbers are memory operations AND processor operations combined and when I say operations I mean quite massive ones. Only writing X MB of data to memory would of course be much faster than writing the same amount X MB of data to disk.
TL;DR You can not compare ProcMem MB/S with disk write MB/s numbers.
@Daniel15 and others
Thanks a lot for the compliments. LETters liking the result of my work is my only "payment" and seeing that it's useful for some here makes me smile. So thank you.
This is so cool. Thanks for taking taking your time out to bring this out.
Very long thread but also very detailed.
I have 16GB AMS and its amazing
For those who mention how long it is: hey, keep in mind that I benchmarked 11 VPSs in 9 locations! That is bound to be lengthy but I tried hard to structure it somewhat so that those of you who are mainly interested in one or two locations only can get at their "meat" quickly.
Yes, thats really amazing work actually. Thanks for benchs
I have two server with them. And my choice is not wrong. Finally, I can rest from burning my money. Hosthatch ftw!!!
@jsg
My ticket 507219 is nearly 10 days old, and no reply has been received.
That's what you're talking about"helpful, and responding quickly"?
I tried to compile Disk ops from OP's post. Disclaimer: Blame OP for any data accuracy issue j/k
TL'DR: Read/Write for both buffered and direct ops are very consistent at high number across all locations. Avg read reaches up to 6.3GB/s, avg write up to 4.7GB/s, very high speed. Coefficient of variation ranges from 10.5% to 24%, with average 15%, meaning good consistency among all types of operations. It's worth to mention that even the performance of Zurich storage box is on par with other NVMe boxes in some operations. Small box is not very impressive in raw number, but still a good choice considering its performance per price ratio.
For buffered read/write among all locations:
For direct read/write among all locations:
For data nerd, here are some stats:
For data paranoid, here are the raw data:
Yes, that was my experience. Whenever I contacted support they knew their stuff and their boxes well and acted and responded quickly.
That is not to mean that there are no exceptions but exceptions and occasional glitches are normal and everywhere, every provider I know has them, e.g. during BF/CM. Plus, of course, I can't judge your case as I do not know the context and what your ticket was about.
Small addendum
I just closed the file and saw that it actually were over 1500 result sets.
Other stuff:
When cleaning up after myself I noticed that the @hosthatch panel regrettably always shows defaults or stuff like "select ISO". I think it would be highly desirable to rather show the currently active value/selection, e.g. the currently selected boot order.
The average write rate during dd'ing the partition I had used was about 550 MB/s. Not bad at all. In fact remarkably good, but then it's not new that HostHatch's disks are insanely fast
When I started set up the 11 VPSs I was about to benchmark about a week ago bootin and rebooting a VPS would sometimes be quite slow. HostHatch's support explained that that was due to BF/CM crazy sales volume. And indeed when I now rebooted it was very fast (as in "a couple of seconds").
All in all/TL;DR: A really great experience. I sincerely can warmly recommend HostHatch VPSs.
Happy new year to everyone - may it be a better one than 2020.
Thank you for the detailed numbers and discussion around them @jsg! I have actually commented the very same previously, that @HostHatch performance is boringly consistent.
Abdullah did put emphasis on using older hardware for ensuring predictable performance over time. So far I do see this on my 3 VPS with them. I really have nothing to complain on their hardware.
And on their network, I agree that M247 is pretty weak overall, but I also see somehow different paths than with most hosts on M247? My experience is mostly with VPNs on M247 and they usually have horrible peering with my ISP. Yet HostHatch works fine and saturates my speeds. So..a better crap?
Happy New Year everyone!
Thanks, and ...
Yes, I was under that impression too. In fact I wondered whether M247 has different quality levels. While I wouldn't use the term "better crap" in the context of @hosthatch I certainly wouldn't call their connectivity great - but clearly better than what one usually sees with M247 connected providers.
Just to make sure, I was ironic and playing with the reputation of M247. HostHatch network, for some reason, works fine for all my uses!
exactly, the current active selection function is highly suggested.
hope hosthatch improve that soon
All of Balkan runs through Vienna. Bratislava runs through Vienna. Much of Hungary runs through Vienna and practically all large Eastern carriers (RU/UA) are present on VIX. My route from Croatia (A1) to Slovenia (Telekom) runs through... Vienna (via DTAG).
Bratislava is (aside of cheaper power, but not much) fairly irrelevant, Milan is useful in some cases but more West EU centered.
They have good sausages