Ξ welcome to cryptostorm's member forums ~ you don't have to be a cryptostorm member to post here Ξ
Ξ any OpenVPN configs found on the forum are likely outdated. For the latest, visit here or GitHub Ξ
Ξ If you're looking for tutorials/guides, check out the new https://cryptostorm.is/#section6 Ξ

The CryptoStorm Speed Test Thread

Post a reply


This question is a means of preventing automated form submissions by spambots.
Smilies
:D :) ;) :( :o :shock: :? 8-) :lol: :x :P :oops: :cry: :evil: :twisted: :roll: :!: :?: :idea: :arrow: :| :mrgreen: :geek: :ugeek: :angel: :clap: :crazy: :eh: :lolno: :problem: :shh: :shifty: :sick: :silent: :think: :thumbdown: :thumbup: :wave: :wtf: :yawn:
BBCode is ON
[img] is ON
[flash] is OFF
[url] is ON
Smilies are ON
Topic review
   

If you wish to attach one or more files enter the details below.

Expand view Topic review: The CryptoStorm Speed Test Thread

Re: The CryptoStorm Speed Test Thread

by KungFuChe » Tue Jan 15, 2019 10:00 pm

I noticed the web site and forum overhaul... so the home page now says: "The OpenVPN service provider for the truly paranoid". Well sure, anyone who is not paranoid is not paying attention ... but does this mean that plans to support wireguard have been abandoned? If not, can you give us some hint about your timeline for trials and full scale deployment? Are we talking weeks or months? Just wanted to have some idea of where I should be sending clients who ask about wireguard service.

Re: The CryptoStorm Speed Test Thread

by blurb » Thu Nov 01, 2018 10:20 am

@df

Yeah, I glossed over everything but the speed comparison there didn't I. Thanks for clarifying the privacy implications, and I agree there isn't any privacy if they chose for there not to be - the power is all with them.

I do find those $5 instances fun to play with though, when I just want to tinker with something for its own sake, and treat them as a 'crowded room', temporary and disposable - probably spend a dollar a month. If any of my vague ideas go anywhere I'll take a good look at using the more solid options you linked, nice one. I liked those prices.

You guys will continue to be the ones to keep the list of the sites my family visit off my ISP's logs, and by extension out of the hands of the ~50 organisations my government said can access them as they please. Not due to it being very interesting in and of itself, non of us would even blush, but due to fuck-them; the principal. I much prefer the attitude I've picked up from you guys so you'll be our quasi isp.
"Also, sometime soon we'll add Wireguard to all our servers."
:clap:

Yay. If you want the perspective of a mostly casual linux user who'll funnel 300gig/month of irritating youtube videos, xbox updates etc through it to beta test, I'd be more than willing to jump on early. I'm looking for an excuse to integrate it into my router and squeeze that last bit out of our connection.

Re: The CryptoStorm Speed Test Thread

by df » Wed Oct 31, 2018 12:20 pm

@blurb
With the Wireguard option it probably would be, but I wouldn't trust a VPS for anything that requires a high
level of security. Since it's a VM, you have no way of verifying the security of the system running the VM. Whatever security measures you do (firewall, grsecurity, FDE, strong passwords/keys, etc.) can all be circumvented if someone has root access to the host server. Plus, VPS bandwidth tends to be fairly limited/expensive in some regions. Probably enough for one person though.

Also, sometime soon we'll add Wireguard to all our servers.
I'd recommend waiting for that, but if you require more immediate traffic obfuscation than OpenVPN's --tls-crypt provides (what our ECC instances use), I'd say go with streisand. That setup does offer more obfuscation methods than we currently do.
But after or during our wireguard implementation, we'll most likely be adding extra obfuscation methods to our setup, so you could just wait for that as well.

FYI, online.net, OVH, oneprovider.com, and kimsufi usually offer very cheap 1gbps dedicated servers in France and/or the Netherlands.
At the moment https://oneprovider.com/dedicated-servers/paris-france has a 2.5gbps one for €11/month, and
https://www.online.net/en/server-dedica ... t-2-s-sata has a 1gbps one for €9.99/month.
With those, you wouldn't have to worry about the security issues that a VPS has since they're dedicated servers.

Re: The CryptoStorm Speed Test Thread

by blurb » Tue Oct 30, 2018 2:11 pm

Question: Is a cheap vps set up using https://github.com/StreisandEffect/streisand faster than Cryptostorm? To make it a bit unfair, I'm connecting directly from my laptop and not the router.

Comparison is to my last post in this thread.

$ ./speedtest-cli
Retrieving speedtest.net configuration...
Testing from Linode (***.***.***.***)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by 23media GmbH (Frankfurt) [0.53 km]: 31.704 ms
Testing download speed.........................................................
Download: 34.58 Mbit/s
Testing upload speed...........................................................
Upload: 8.87 Mbit/s

Answer: Not really, no, not on my old laptop.


Just thought I'd share.

Re: The CryptoStorm Speed Test Thread

by parityboy » Wed Oct 17, 2018 6:48 pm

@blurb

Many thanks for sharing this. :D

Ironically, while the newer ECC cryptographic algorithms are much more efficient than the older RSA - requiring much less CPU power for the same level of cryptographic strength - there's a good chance that the older routers (which would actually benefit from this) will never be updated to take advantage of it.

Re: The CryptoStorm Speed Test Thread

by blurb » Wed Oct 17, 2018 3:13 pm

Post upgrade 'ECC':

VPN connected to ASUS RT-AC68U ('FreshTomato' firmware) overclocked to 1200/800.
OpenVPN 2.4.6, OpenSSL 1.0.2p
Base connection is ~40 Mbits down, 10 Mbits up. VDSL.
Connected to closest exit node.

Test was conducted on an old (and much loved) thinkpad connected to router via ethernet. I'm not bothering with a wifi test, too many variables for it to mean much imho.

$ ./speedtest-cli
Retrieving speedtest.net configuration...
Testing from UK Dedicated Servers Limited (5.101.149.4)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by Uno (Milton Keynes) [49.22 km]: 14.871 ms
Testing download speed................................................................................
Download: 33.77 Mbit/s
Testing upload speed................................................................................................
Upload: 8.95 Mbit/s

$ ./speedtest-cli
Retrieving speedtest.net configuration...
Testing from UK Dedicated Servers Limited (5.101.149.4)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by Uno (Milton Keynes) [49.22 km]: 15.385 ms
Testing download speed................................................................................
Download: 34.43 Mbit/s
Testing upload speed................................................................................................
Upload: 8.93 Mbit/s

$ ./speedtest-cli
Retrieving speedtest.net configuration...
Testing from UK Dedicated Servers Limited (5.101.149.4)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by Uno (Milton Keynes) [49.22 km]: 15.958 ms
Testing download speed................................................................................
Download: 34.74 Mbit/s
Testing upload speed................................................................................................
Upload: 8.86 Mbit/s


I thought the new fancy ways would be a bit slower, but it seems not. :thumbup:

Re: The CryptoStorm Speed Test Thread

by parityboy » Fri Jun 02, 2017 8:23 am

Connection: 4G
Exit Node: England

Code: Select all

wget http://cachefly.cachefly.net/100mb.test
--2017-06-02 04:19:36--  http://cachefly.cachefly.net/100mb.test
Resolving cachefly.cachefly.net (cachefly.cachefly.net)... 205.234.175.175
Connecting to cachefly.cachefly.net (cachefly.cachefly.net)|205.234.175.175|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 104857600 (100M) [application/octet-stream]
Saving to: ‘100mb.test.2’

100%[================================================================================================>] 104,857,600 3.36MB/s   in 31s    

2017-06-02 04:20:12 (3.25 MB/s) - ‘100mb.test.2’ saved [104857600/104857600]

Re: The CryptoStorm Speed Test Thread

by frankfooter » Sun May 07, 2017 6:23 am

25mbit DSL
US North Node
Downloading http://cachefly.cachefly.net/100mb.test
900KB/sec

Re: The CryptoStorm Speed Test Thread

by parityboy » Sat Apr 08, 2017 6:58 pm

@JTD121

Nice to see that the upstream throughput is inline with your "naked" package performance. Assuming the VPN connection starts on your PC, the downstream throughput could possibly point to congestion on that node.

Is that node the closest one to your location?

Re: The CryptoStorm Speed Test Thread

by JTD121 » Thu Mar 30, 2017 4:59 pm

Image

Just messin' around, and used DSLReports speed test! 60s test time, 32 download/upload streams (not sure how this affects the test).

I believe this was connected to the US North or US East node cluster....

My speed tier is 200mbps down, 10mbps up. Using an Arris SB6190 with the Puma6 issues.

Re: The CryptoStorm Speed Test Thread

by Russell » Wed Mar 22, 2017 12:33 am

I guess it's my turn now.

How connection is being made: Duplex FTTB 25Mb/25Mb cable with cheap TP-LINK router (very stable commection. Ping between three local cities is only 9ms).
Exit nodes: Using Swiss and German exit nodes for testing.
OS: Windows 7.

Speedtest (ping/dl speed/upload speed): Original - 3ms/11.85Mbs/11.33Mbs (it's evening time, switch is probably overloaded)
Swiss node (My provider isn't hyper-restrictive, so I mainly use UDP) - 71ms/4.50Mbs/7.39Mbs (pretty decent result considering the fact that Swiss isn't that close to me)
German node (Same story - UDP, Frankfurt) - 92ms/2.92Mbs/8.87Mbs (it's LeaseWeb, so no wonder)
German node (UDP, Dusseldorf) - 97/3.8Mbs/5.04Mbs (Nothing unusual, typical German node).

I only use these as my daily driver because they're located FAIRLY close to me.

Yesterday I've had 52ms for ping and 10Mb/s for both dl/upload at Swiss node.

I usually get 1.4MB/s during torrent download. That's more than enough for me. If I'll need high speed for multimedia, I have a dedicated line and router for that.
 ! Message from: parityboy
Edit throughput units for clarity.

Re: The CryptoStorm Speed Test Thread

by cryptocornhole » Thu Feb 02, 2017 11:02 pm

Here are my results:

Code: Select all

wget -O /dev/null http://speedtest.ams01.softlayer.com/downloads/test100.zip
No VPN:

Code: Select all

2017-02-02 10:45:25 (3.65 MB/s) - ‘/dev/null’ saved [104874307/104874307]
cstorm_USsouth:

Code: Select all

2017-02-02 10:44:39 (1.07 MB/s) - ‘/dev/null’ saved [104874307/104874307]
cstorm_USnorth:

Code: Select all

2017-02-02 10:50:38 (1.67 MB/s) - ‘/dev/null’ saved [104874307/104874307]
cstorm_USeast:

Code: Select all

2017-02-02 10:52:59 (1.85 MB/s) - ‘/dev/null’ saved [104874307/104874307]
cstorm_UScentral:

Code: Select all

2017-02-02 10:55:52 (812 KB/s) - ‘/dev/null’ saved [104874307/104874307]

Re: The CryptoStorm Speed Test Thread

by tenaciousj » Tue Nov 29, 2016 12:18 pm

I ran several tests from Charlotte, NC and got pretty good results on a
300Mbps cable connection routed over 802.11ac to the client
I normally get 150-200Mbps wifi speeds, so pulling 100Mbps on speedtest over VPN is pretty damn fast

large file wget download results:
(connected to the std. US East connection)

100mb file:
atlanta 3.71MB/s time 30s
newark 7.41MB/s time 13s
dallas 4.74MB/s time 21s

1000mb from Leaseweb in Virginia gave me 5.26MB/s time: 3m 1s

speedtest.net (for kicks):
100Mbps down 16.42Mbps up (up is capped at 20)

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Sun Nov 20, 2016 1:11 pm

Switzerland server, this time from the command line with speedtest-cli.

Code: Select all

Retrieving speedtest.net configuration...
Testing from DataSource (XXX.XX.XXX.XX)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by Datasource. AG (Zug) [6.26 km]: 51.064 ms
Testing download speed................................................................................
Download: 78.53 Mbit/s
Testing upload speed....................................................................................................
Upload: 39.13 Mbit/s

Re: The CryptoStorm Speed Test Thread

by parityboy » Fri Oct 07, 2016 1:39 am

@thisis4thepeople

I was tempted to rent a 1Gb/s dedi and install a desktop on it just to get some numbers for my own tests, but I see there's no need. CS nodes are certainly fast enough to support the user base. :D

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Thu Oct 06, 2016 5:48 pm

The Switzerland server is still speedy:

Image

Re: The CryptoStorm Speed Test Thread

by Khariz » Tue Sep 13, 2016 11:03 am

My speeds are great. I just tested all the U.S. Servers and England and I'm getting 45down, 35up on my 50/50 connection.

I'm even getting 30down on the Isle of Man voodoo.

Re: The CryptoStorm Speed Test Thread

by wpaschukat » Sun Sep 11, 2016 6:03 pm

Might as well..

Paris node, on a 100/10 line:

Code: Select all

wget -O /dev/null http://speedtest.ams01.softlayer.com/downloads/test100.zip
HTTP request sent, awaiting response... 200 OK
Length: 104874307 (100M) [application/zip]
Saving to: ‘/dev/null’

/dev/null                                    100%[==============================================================================================>] 100.02M  6.42MB/s   in 17s
More than reasonable :)
 ! Message from: parityboy
Edited to improve formatting

Re: The CryptoStorm Speed Test Thread

by parityboy » Sun Sep 11, 2016 5:07 pm

@sm1th

Many thanks for sharing this. Your numbers look in line with what you connection can support. :)

Re: The CryptoStorm Speed Test Thread

by sm1th » Wed Sep 07, 2016 2:04 pm

Ok, I'll bite.

40Mb bareback connection, over 5ghz wifi via Asus RT-AC68U (overclocked to 1200/800). England node. I've used the speedtest file from a post above...for some vague semblance of comparison (or I'm just lazy).

Code: Select all

$ wget -O /dev/null http://speedtest.ams01.softlayer.com/downloads/test100.zip

100%[======================================================================================================================================>] 104,874,307 2.92MB/s   in 36s    

2016-09-07 09:50:29 (2.77 MB/s) - ‘/dev/null’ saved [104874307/104874307]

Re: The CryptoStorm Speed Test Thread

by parityboy » Mon Sep 05, 2016 12:02 am

@thisis4thepeople

Many thanks for keeping us up-to-date with throughput results, it's greatly appreciated. :D If only others would chime in...

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Fri Sep 02, 2016 2:27 pm

I'm back with new results. Last time I had a down-stream of over 13 MB/s, now it's the other way around:

Image

Not to shabby. This was via Denmark by the way.

Re: The CryptoStorm Speed Test Thread

by parityboy » Tue Jul 12, 2016 3:11 am

@turbz

Many thanks for your post, it's nice to see people still contributing to this thread. Proves that secure VPNs don't have to be slow. :D

Re: The CryptoStorm Speed Test Thread

by turbz » Tue Jul 12, 2016 12:06 am

Well, finally I get acceptable speeds :)
Constant +18MBps in and +2MBps out, this is on 350/30 line.
Switzerland node.

Re: The CryptoStorm Speed Test Thread

by parityboy » Wed Jun 29, 2016 5:23 am

@thisis4thepeople

Many thanks for keeping us updated with your results, it's greatly appreciated. Unfortunately, I'm still stuck on an 8/1 ADSL line, so I'm unable to contribute further.

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Sun Jun 19, 2016 6:16 pm

New day, new record (Switzerland server): 13.26 MB/s equates to over 100 Mbit/s (106 to be precise). This is on a 200 Mbit/s line if I'm not mistaken. My ISP gave me 100 Mbit/s before but I think it has been upgraded in the meantime.

Image

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Fri May 06, 2016 5:50 pm

Managed a new record today, I wasn't even paying attention when I heard the fans of my device go on and checked to see what was going on. To my surprise, I was uploading at quite a nice speed, around 75 Mbps.

Connected to the Oslo node by the way.

Image

Re: The CryptoStorm Speed Test Thread

by cumpsd » Sun Apr 03, 2016 1:30 am

Strange results...

Testing by downloading http://speedtest.ams01.softlayer.com/do ... est100.zip via wget

Without VPN / Belgium:
240Mbps down / 20 Mbps up
2016-04-02 00:56:00 (25.6 MB/s) - 'test100.zip' saved [104874307/104874307]

Paris UDP:
2016-04-02 01:03:00 (259 KB/s) - 'test100.zip.1' saved [104874307/104874307]
Paris TCP:
2016-04-02 02:00:22 (2.06 MB/s) - 'test100.zip.11' saved [104874307/104874307]

Frankfurt UDP:
2016-04-02 01:18:18 (269 KB/s) - 'test100.zip.3' saved [104874307/104874307]
Frankfurt TCP:
2016-04-02 01:58:56 (1.37 MB/s) - 'test100.zip.10' saved [104874307/104874307]

Switzerland UDP:
2016-04-02 01:32:14 (267 KB/s) - 'test100.zip.5' saved [104874307/104874307]
Switzerland TCP:
2016-04-02 02:02:35 (7.21 MB/s) - 'test100.zip.12' saved [104874307/104874307]

Lisbon UDP:
2016-04-02 01:11:07 (233 KB/s) - 'test100.zip.2' saved [104874307/104874307]

Rome UDP:
2016-04-02 01:25:20 (277 KB/s) - 'test100.zip.4' saved [104874307/104874307]

Voodoo-USWest-Iceland:
2016-04-02 01:46:17 (287 KB/s) - Read error at byte 103115450/104874307 (Invalid argument). Retrying.

Voodoo-USSouth-IsleOfMan:
No results, kept reconnecting and was very unstable

Re: The CryptoStorm Speed Test Thread

by speedtesting » Sat Jan 16, 2016 5:29 pm

Bare connection = 40Mbp/s by provider's router in modem only mode.
Local = ethernet cable - netgear switch - 500Mbp/s powerline - ASUS RT-AC68R/U (Shibby Tomato 132 running OpenVPN. Client 2 being used for this test).

Tested from UK.

----------

London.

$ wget -O /dev/null http://releases.ubuntu.com/14.04.3/ubun ... -amd64.iso
--2016-01-16 11:44:00

100%[============================================================================================================================================================>] 601,882,624 2.67MB/s in 3m 20s

2016-01-16 11:47:20 (2.87 MB/s) - ‘/dev/null’ saved [601882624/601882624]

-----------

Paris.

$ wget -O /dev/null http://releases.ubuntu.com/14.04.3/ubun ... -amd64.iso
--2016-01-16 11:48:36--

100%[============================================================================================================================================================>] 601,882,624 2.79MB/s in 4m 55s

2016-01-16 11:53:32 (1.94 MB/s) - ‘/dev/null’ saved [601882624/601882624]


----------

Frankfurt.

$ wget -O /dev/null http://releases.ubuntu.com/14.04.3/ubun ... -amd64.iso
--2016-01-16 11:56:12--

100%[============================================================================================================================================================>] 601,882,624 2.22MB/s in 4m 38s

2016-01-16 12:00:50 (2.06 MB/s) - ‘/dev/null’ saved [601882624/601882624]

-----------

Switzerland.

*$ wget -O /dev/null http://releases.ubuntu.com/14.04.3/ubun ... -amd64.iso
--2016-01-16 12:02:47--

100%[============================================================================================================================================================>] 601,882,624 2.95MB/s in 5m 39s

2016-01-16 12:08:26 (1.69 MB/s) - ‘/dev/null’ saved [601882624/601882624]

Re: The CryptoStorm Speed Test Thread

by parityboy » Fri Jan 01, 2016 10:31 pm

@thread

Internet Connection: Cable, 50Mb/s
Local Connection: 802.11n, 72Mb/s as tested

Code: Select all

Node           wget       Torrent (avg)    Torrent (peak)
Naked          5.10MB/s   4.35MiB/s        4.99MiB/s
Cantus         3.74MB/s   3.76MiB/s        5.10MiB/s
Jord           2.60MB/s   3.57MiB/s        4.43MiB/s
Tagus          2.63MB/s   3.15MiB/s        3.48MiB/s
Turing         4.58MB/s   3.88MiB/s        4.53MiB/s

Re: The CryptoStorm Speed Test Thread

by vehf277 » Wed Dec 02, 2015 9:42 pm

What about testmy.net and speedof.me for speed test usage?

Re: The CryptoStorm Speed Test Thread

by JTD121 » Wed Nov 25, 2015 9:20 am

Still can't seem to edit my posts above, but this is Test 4, connected to Buffalo, NY node node.
  • Test 4: GOG Downloader; downloading ~830MB in total
  • Speeds don't seem to fluctuate as in previous tests
  • Speeds from 650KB/sec to 1.4-1.5MB/sec; Very consistent staying towards the top of this range
  • Prolonged idle (reading a single page of something) doesn't seem to affect overall bandwidth.
  • Also running a multi-threaded test at http://testmy.net for extra bandwidth loading 8-)
  • Not really using all bandwidth available to me without CS VPN running, but definitely getting better!
  • No DI.FM streaming this time

Re: The CryptoStorm Speed Test Thread

by JTD121 » Sat Nov 21, 2015 6:27 pm

This test site was suggested in IRC, and it looks like it's not Flash-based. Also, has an adjustable long-term speed testing method. Currently testing it out without CS to see how it fares.

Will obviously do some stress testing and compare to CS as well.

Re: The CryptoStorm Speed Test Thread

by JTD121 » Sat Nov 21, 2015 2:24 am

Hm...My connection is rated about 25/5, and I've regularly seen speeds around 31/7, so I'll give this a couple test downloads soon.

I'll try to keep similar format as before (again borrowing marzametal's format).

Re: The CryptoStorm Speed Test Thread

by parityboy » Thu Nov 19, 2015 11:34 pm

@JTD121

Might be an idea. My connection is too slow to make any difference, but for those with 30Mb+ connections, it might prove useful. :)

EDIT:
Taken from this post:

Code: Select all

france:                Speed: 1000Mb/s
frankfurt:             Speed: 100Mb/s
london:                Speed: 1000Mb/s
france#2:              Speed: 1000Mb/s
lisbon:                Speed: 1000Mb/s
seattle:               Speed: 1000Mb/s
singapore:             Speed: 100Mb/s
switzerland:           Speed: 1000Mb/s
italy:                 Speed: 100Mb/s
vegas:                 Speed: 1000Mb/s
canada:                Speed: 100Mb/s
moldova:               Speed: 100Mb/s

Re: The CryptoStorm Speed Test Thread

by JTD121 » Thu Nov 19, 2015 10:30 pm

So with the new nodes and replacement nodes up and running, and using gigabit pipes, do we want to retest some of them?

Re: The CryptoStorm Speed Test Thread

by JTD121 » Sun Oct 18, 2015 5:22 pm

Can't seem to edit my post above, but this is Test 3, connected to London - England node.
  • Test 3: Univ of Kent UK (geographically close to London node (I assume))
  • Speeds don't seem to drop as low as previous US-central node
  • Speeds from 75KB/sec to 1.5MB/sec; mostly seem to fluctuate in the 150 ~ 450KB/sec range
  • Prolonged idle (reading a single page of something) seems to give download a bit of a boost; observed increases between 600KB/s to 750KB/sec, several bursts 30~40 second to 1.3 - 1.5MB/sec
  • Same download setup as before, just different node.
  • DI.FM stream didn't skip a beat (har har!) during download

Re: The CryptoStorm Speed Test Thread

by parityboy » Tue Oct 13, 2015 4:00 am

@rwolslegel,@JTD121,@DonderPjem

Many thanks for posting your experiences in this thread. It would appear that routing across the United States is awful, at least for the routes used by the datacentres where Cryptostorm nodes are located. The paths to Switzerland aren't too hot either, by all accounts.

Re: speedtest.net, yes it's bursty, also browsers running a speed test written in Flash(!) aren't the most reliable things in the world. File transfer software such as FileZilla or wget (for the 'nix heads out there) gives far more accurate results.

Re: The CryptoStorm Speed Test Thread

by rwolslegel » Tue Oct 13, 2015 2:04 am

I spent lots of time running tests and have generated a ton of data. Here it is:

All tests were done within about 30 minutes of each other using wget with five different servers and three different hosting providers. All were done from GNU/Linux. I am physically located on the eastern side of the United States.

Softlayer - Dallas US:
  • Not Connected to CS: 29.2MB/s
  • US Central: 3.03MB/s
  • US West: 1.2MB/s
  • Montreal: 6.90MB/s
  • London: 7.23MB/s
  • Switzerland: 1.04MB/s

Softlayer - Seattle US:
  • Not Connected to CS: 20.3MB/s
  • US Central: 2.17MB/s
  • US West: 2.11MB/s
  • Montreal: 7.64MB/s
  • London: 5.85MB/s
  • Switzerland: 1019KB/s
Softlayer - Amsterdam NL:
  • Not Connected to CS: 19.0MB/s
  • US Central: 2.52MB/s
  • US West: 632KB/s
  • Montreal: 4.6MB/s
  • London: 10.8MB/s
  • Switzerland: 1.2MB/s
Linode - Atlanta US:
  • Not Connected to CS: 24.3MB/s
  • US Central: 1.78MB/s
  • US West: 1.74MB/s
  • Montreal: 5.47MB/s
  • London: 7.01MB/s
  • Switzerland: 434KB/s
DigitalOcean - New York US (A VPS that I spun up just to experiment):
  • Not Connected to CS: 28.3MB/s
  • US Central: 1.73MB/s
  • US West: 1.67MB/s
  • Montreal: 5.84MB/s
  • London: 5.97MB/s
  • Switzerland: 230KB/s
Some conclusions that I can draw from this is that the fastest node for me is either Montreal or London, while the node that's actually closest to me is one of the slowest (uscentral). This is consistent with other observations I have made in the past, for example being unable to stream 1080p Youtube when connected to US nodes.

Re: The CryptoStorm Speed Test Thread

by JTD121 » Tue Oct 13, 2015 12:49 am

To borrow/adapt marzametal's format:

How connection is made:
  • Comcast cable; rated for 25/5, though real-world it's 30/6 on-the-wire (isn't Comcast generous?)
  • Asus RT-N66U running Advanced Tomato 2.9-K131
  • Wireless desktop, using 5GHz 802.11N, reads full 450Mbps connection
  • Test 1: Kansas City - US Central node
  • Test 2: Toyko - Japan (because Japan :D )
  • Windows Widget v2.22: I believe random UDP port; currently 41217. All 3 in Security tab checked
  • Single download at a time (one HTTP download, one torrent, etc)
Download methods: Linux Mint 17.2 Cinnamon 64-bit (using KB = kilobyte, kb= kilobit as reported by software)
  • Test 1: From University of Oklahoma mirror
  • Also streaming from Digitally Imported at 128k AAC during tests (so ~130-150 kb/s with overhead); loves me some EDM :D :thumbup:
  • Download via Firefox 41.0.1 between 31.2KB/s (absolute minimum observed) and 499KB/s (absolute maximum observed); also notice a dip in speed when loading up this thread, or other small (non-heavy interactive/download) sites (Slashdot (though not too light anymore), Twitter, Gmail (Ctrl+F5 refresh)
  • Speeds seem to hover in the 100-250KB/sec range, jumping up and down every so often.
  • Test 2: Same mirror (Univ of OK)
  • Speeds as low as 5.5KB/sec, almost never getting higher than 60KB/sec
  • Talk about stress-test, huh?
  • Same download setup as before, just different node.
  • Made DI.FM stream drop out repeatedly, while also affecting the download speed of the ISO
Disconnecting from CS, I see that same download, from the same mirror, it's superluminal; pegging the connection at 2.5-3.6MB/sec (hovering at 2.8-3.4MB/sec). This information is to help the network, so please don't take it as a detrimental post I've created here. Not sure the exact setup for connectivity at the node, within your kernels, etc. Just for point of reference.

Will update with additional information as I run tests (torrents, different nodes, different mirrors)

Re: The CryptoStorm Speed Test Thread

by JTD121 » Mon Oct 12, 2015 9:16 pm

I definitely want to help out; couple questions. Why do we avoid speedtest.net? Probably because it's a bursty test, rather than an average or median test?

Are there any particular software to track this on the user end? I know uTorrent has a sort of speed tracker over the download portion of a torrent, but outside of torrents?

Currently using the eyeball method while watching an HTTP download from a mirror.

Re: The CryptoStorm Speed Test Thread

by DonderPjem » Sun Oct 11, 2015 12:45 am

Thanks for your time!
Ive tried every mirror with the respective exit node and note down the maximum speed it maintained for about 10 seconds in the first 2 minutes.

(speeds are in KB/s and mirrors are in order of appearance on http://www.linuxmint.com/edition.php?id=172)

france
300
250
400
250
600

germany
300
900
200
350
300
700
600

uk
400
offline

switzerland
550

netherlands (native connection)
3000
2200

netherlands (germany exit node)
500
350

Re: The CryptoStorm Speed Test Thread

by parityboy » Sat Oct 10, 2015 4:44 pm

@DonderPjem

Thanks for the info. :)

Would it be possible for you to conduct a couple of tests using the links at the top of the page? Basically, connect to an exit node closest to your location, then pick a file source for Linux Mint which is also closest to your location - so for example if you are in Germany, you would use the Frankfurt exit node and download the ISO from this link.

If you could report back with the results of a couple of runs, it would be appreciated. :)

Re: The CryptoStorm Speed Test Thread

by DonderPjem » Fri Oct 09, 2015 9:57 pm

I'm using the cryptostorm widget on my 64bit windows machine with an i7 2630QM. The cpu usage is insignificant. I'm pretty sure the bottleneck is not on my part. PIA gave me speeds at around 95% of my native bandwidth. Ive tried every server cluster by now and I havent been able to utilize more than 15% of my 100mbit connection yet.

Re: The CryptoStorm Speed Test Thread

by parityboy » Wed Oct 07, 2015 10:39 pm

@DonderPjem

Can you give a little more details as to your setup?

1. Is the VPN endpoint on your router or on your desktop? What's the respective CPU load?
2. What CPU/OS are you running?

Thanks. :)

Re: The CryptoStorm Speed Test Thread

by DonderPjem » Wed Oct 07, 2015 5:44 pm

Im in the Netherlands and the london, frankfurt and paris clusters all give the same performance.

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Wed Oct 07, 2015 11:02 am

Preferably, you'd choose a server close to you. Test all servers closest to you and use the one that gives you the highest speed, if that is what you prefer.

Re: The CryptoStorm Speed Test Thread

by DonderPjem » Wed Oct 07, 2015 8:28 am

My tradeoff for using CS is almost 90% of my bandwidth. My usually 100mbit up/down is reduced to 10~12mbit. Fortunately I only have to endure this during internet browsing as I outsource my heavy torrenting to a seedbox. But it still becomes noticable when trying to watch yt videos in 1440p or download something larger than a few megs.

It looks like I am being hit the hardest here, does anyone have optimalisation tips? Should I just hope for a server cluster in my country?

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Tue Oct 06, 2015 1:38 pm

Image

64 Mbit/s on UK server on a 100 Mbit/s line.

Re: The CryptoStorm Speed Test Thread

by parityboy » Tue Aug 04, 2015 4:59 am

@sysfu

From the OP
2. DO NOT use speedtest.net. Since we're trying to measure real world throughput, please use a real world application, such as a web browser doing an HTTP download, BitTorrent client, FTP client (FileZilla) or even curl or wget.

Re: The CryptoStorm Speed Test Thread

by sysfu » Sun Aug 02, 2015 11:03 pm

Using exit linux-cantus1.xn--cdaan2d.be46.165.222.248
Native 22Mbps/2Mbps dn/up

Code: Select all

Retrieving speedtest.net configuration...
Retrieving speedtest.net server list...
Testing from Leaseweb Germany GmbH (46.165.222.248)...
Selecting best server based on latency...
Hosted by Base-Mail (Frankfurt) [100.73 km]: 237.964 ms
Testing download speed........................................
Download: 8.62 Mbit/s
Testing upload speed..................................................
Upload: 0.98 Mbit/s

Re: The CryptoStorm Speed Test Thread

by parityboy » Thu Jul 09, 2015 4:18 am

Thracky wrote:
parityboy wrote:@Thracky

On your faster speed runs via CS, what was your CPU load like?
One core bouncing around between 9-25% on the openvpn process. (i7-2600k, no overclock)
Well, 25% load tells me the issue certainly isn't at your end. I'm beginning to wonder if the issue is the OpenVPN process itself, or maybe further down the networking stack? Considering the fact that Linux-based routers are capable of throughput far greater than 50Mbit/s - combined with the fact that CS' exit nodes are full-blown x86 servers - I'd say the throughput issue lies with OpenVPN itself

However, without proper testing I'm just guessing. :)

Re: The CryptoStorm Speed Test Thread

by Thracky » Thu Jul 09, 2015 2:33 am

parityboy wrote:@Thracky

On your faster speed runs via CS, what was your CPU load like?
One core bouncing around between 9-25% on the openvpn process. (i7-2600k, no overclock)

Re: The CryptoStorm Speed Test Thread

by parityboy » Thu Jul 09, 2015 12:16 am

@Thracky

On your faster speed runs via CS, what was your CPU load like?

Re: The CryptoStorm Speed Test Thread

by marzametal » Wed Jul 08, 2015 8:06 am

Thracky wrote: US Central - 142.54.172.51
Seriously, this was going between 5-20KB/sec if at all so I cancelled it.
This one was a pain for me this morning... the others seem to serve me well.

Re: The CryptoStorm Speed Test Thread

by Thracky » Wed Jul 08, 2015 6:15 am

So I'm in my first few days of cryptostorm and it's pretty painful for me so far. Here's why.

Connection Details:
250 Mb down/20 Mb up - Eastern-ish Canada

All tests performed using wget of http://speedtest.dal01.softlayer.com/do ... est100.zip

No VPN

Code: Select all

100%[===============================================>] 104,874,307 21.4MB/s   in 8.4s   

2015-07-07 20:27:34 (11.9 MB/s) - ‘/dev/null’ saved [104874307/104874307]
File wasn't completely large enough to show my normal sustained speed but the peak at the end is pretty typical and even a bit on the low side.

Montreal Exit - 198.27.76.1

Code: Select all

2015-07-07 20:53:34 (77.0 KB/s) - ‘/dev/null’ saved [104874307/104874307]
So that was pretty awful. Also, I couldn't even keep a persistent SSH or IRC session without constant disconnects. So, let's try a completely different exit...

US Central - 142.54.172.51
Seriously, this was going between 5-20KB/sec if at all so I cancelled it.

NEXT. Let's see if going across the ocean helps...

Iceland - 109.71.42.163

Code: Select all

2015-07-07 21:03:35 (1.93 MB/s) - ‘/dev/null’ saved [104874307/104874307]
Well that was substantially better and it finished strong at just over 4 MB/s. Let's try one more.

London - 130.180.201.117

Code: Select all

2015-07-07 21:01:30 (4.73 MB/s) - ‘/dev/null’ saved [104874307/104874307]
That's more like it.

For the sake of comparison, let's check a couple exit nodes on PIA:
PIA - Toronto (Closest geographically)

Code: Select all

2015-07-07 21:05:43 (3.32 MB/s) - ‘/dev/null’ saved [104874307/104874307]
PIA - US East

Code: Select all

2015-07-07 21:06:54 (2.95 MB/s) - ‘/dev/null’ saved [104874307/104874307]
And one overseas for good measure, let's try London.

PIA - London

Code: Select all

2015-07-07 21:10:05 (1.74 MB/s) - ‘/dev/null’ saved [104874307/104874307]
And that's that.

So, I'm not particularly thrilled with the North American exits that I tried, but at least Iceland and London performed alright.

Re: The CryptoStorm Speed Test Thread

by bricky149 » Fri Jul 03, 2015 1:22 pm

(Can't edit my posts since the forum upgrade so I'll make a new one.)

I've had to switch to a wired connection, with a 20ft Ethernet cable, because my wireless adapters either drop after 5 minutes or cause Windows to malfunction. On torrents I see ~5MiB/s for download and for regular files it's ~6MiB/s. The highest speed I can achieve on a naked connection is ~14MiB/s.

Re: The CryptoStorm Speed Test Thread

by bricky149 » Thu Jun 18, 2015 2:33 pm

I have a 152Mbps connection at home although it peaks short of 120Mbps IRL. While I'm connected via CStorm I see a top speed of around 50Mbps. The average speed for torrents is around 2.4MiB/s (unchanged from before my connection upgrade) and for regular files I've seen it peak at around 4MiB/s. I've been messaging through Twitter since it's seen as an edge case, difficult to tune my OS since Windows makes it hard to.

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Thu Jun 18, 2015 10:24 am

Another test: One only on YouTube, the other two simultaneous downloads (YouTube + savefrom.net).

67 Mbps on YouTube = 8,375 MB/s
6,8 MB + 16 Mbps = 8,8 MB/s
Attachments
multidl.jpg
ytdl.JPG
ytdl.JPG (27.3 KiB) Viewed 149280 times

Re: The CryptoStorm Speed Test Thread

by sysfu » Tue Jun 16, 2015 9:51 pm

Thought I'd share a pretty cool command line client for speedtest.net that I found the other day: https://github.com/zpeters/speedtest

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Sun Jun 14, 2015 9:45 am

parityboy wrote:
thisis4thepeople wrote:

Code: Select all

curl -o /dev/null http://mirror.nl.leaseweb.net/speedtest/100mb.bin
Exit node: Laika

<snip>
What's your "naked" speed?
100/100 Mbps.

Re: The CryptoStorm Speed Test Thread

by sysfu » Sun Jun 14, 2015 9:33 am

Raw or nekkid connection speed: ADSL ~22Mbps/2Mbps

Re: The CryptoStorm Speed Test Thread

by parityboy » Sun Jun 14, 2015 2:13 am

thisis4thepeople wrote:

Code: Select all

curl -o /dev/null http://mirror.nl.leaseweb.net/speedtest/100mb.bin
Exit node: Laika

<snip>
What's your "naked" speed?

Re: The CryptoStorm Speed Test Thread

by sysfu » Sat Jun 13, 2015 11:13 pm

Direct via ISP

Code: Select all

aria2c http://speedtest.sea01.softlayer.com/downloads/test100.zip

Download Results:
gid   |stat|avg speed  |path/URI
======+====+===========+=======================================================
ed2189|OK  |   2.6MiB/s|D:/Users/test/Downloads/test100.zip
OpenBSD/Alix router via Seattle cluster:

Code: Select all

Download Results:
gid   |stat|avg speed  |path/URI
======+====+===========+=======================================================
a40207|OK  |   415KiB/s|D:/Users/test/Downloads/test100.zip
Narwhal 2.22 widget via Seattle cluster:

Code: Select all

Download Results:
gid   |stat|avg speed  |path/URI
======+====+===========+=======================================================
463ae2|OK  |   1.1MiB/s|D:/Users/test/Downloads/test100.zip

Re: The CryptoStorm Speed Test Thread

by thisis4thepeople » Sat Jun 13, 2015 7:50 pm

Code: Select all

curl -o /dev/null http://mirror.nl.leaseweb.net/speedtest/100mb.bin
Exit node: Laika

Code: Select all

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 95.3M  100 95.3M    0     0  6208k      0  0:00:15  0:00:15 --:--:-- 7315k
Managed to get around 70 Mbps via a public torrent and 50 on YouTube.
Attachments
torrent.jpg
torrent.jpg (16.84 KiB) Viewed 149407 times
st-peter-speedtest-yt.JPG
st-peter-speedtest-yt.JPG (25.93 KiB) Viewed 149407 times

Re: The CryptoStorm Speed Test Thread

by parityboy » Wed Jun 10, 2015 10:03 pm

@sysfu

Ahh, yes I forgot the edit option becomes unavailable after a certain amount of time. No problem. :)

Re: The CryptoStorm Speed Test Thread

by sysfu » Wed Jun 10, 2015 9:45 pm

parityboy wrote:@sysfu

Could you edit your post to include the supported bandwidth (8Mb/s, 24Mb/s, 100Mb/s) of your Internet connection? Many thanks. :)
Forgive me if I'm being dense but I no longer see an option to edit the previous post even though I am logged into the forum.

The 'raw' speed of the Internet connection is ~22Mps down and ~2Mps up. Currently running running OpenVPN + OpenBSD 5.7 on an Alix device which serves entire home LAN.

I just tried another quick and dirty download test from a Windows workstation (with a youtube vid playing mind you) and the results were much better.

Code: Select all

aria2c http://speedtest.dal01.softlayer.com/downloads/test100.zip

Download Results:
gid   |stat|avg speed  |path/URI
======+====+===========+=======================================================
4735ca|OK  |   341KiB/s|C:/Users/test/Downloads/test100.zip
OpenVPN cpu utilization on the Alix hovered around 45% during the download. The previous test was done on a Mac running Yosemite.

I'll also try tearing down the Cryptostorm OpenVPN tunnel on the router and connect directly from both Windows and Mac systems using the current Narwhal widget and Viscosity respectively, to see what kind of results I get.

Re: The CryptoStorm Speed Test Thread

by Fermi » Wed Jun 10, 2015 9:07 pm

Code: Select all

curl -o /dev/null http://mirror.nl.leaseweb.net/speedtest/100mb.bin
Leaseweb Amsterdam

Connection: 200Mb/20Mb Cable

without VPN

Code: Select all

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 95.3M  100 95.3M    0     0  21.2M      0  0:00:04  0:00:04 --:--:-- 21.2M
Exit Node: Cantus (Frankfurt)

Code: Select all

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 95.3M  100 95.3M    0     0  7347k      0  0:00:13  0:00:13 --:--:-- 6977k
Exit Node: Onyx (Paris)

Code: Select all

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 95.3M  100 95.3M    0     0   487k      0  0:03:20  0:03:20 --:--:--  582k
Exit Node: Turing (London)

Code: Select all

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 95.3M  100 95.3M    0     0  5026k      0  0:00:19  0:00:19 --:--:-- 3873k
/Fermi

Re: The CryptoStorm Speed Test Thread

by parityboy » Wed Jun 10, 2015 4:44 pm

@sysfu

Could you edit your post to include the supported bandwidth (8Mb/s, 24Mb/s, 100Mb/s) of your Internet connection? Many thanks. :)

Re: The CryptoStorm Speed Test Thread

by sysfu » Wed Jun 10, 2015 7:45 am

Ran a test today from the Kansas City node to linode test file in Dallas.

¡Qué miserables!

Connection: ADSL 22Mb/2Mb

Code: Select all

curl -o /dev/null http://speedtest.dal01.softlayer.com/downloads/test100.zip

Code: Select all

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  100M  100  100M    0     0  47070      0  0:37:08  0:37:08 --:--:-- 67877
 ! Message from: parityboy
Edited to improve layout.

Re: The CryptoStorm Speed Test Thread

by turbz » Tue Jun 09, 2015 11:58 pm

Thanks a lot for the quick answers and the effort :)

Re: The CryptoStorm Speed Test Thread

by parityboy » Tue Jun 09, 2015 10:13 pm

@turbz

Thanks for that. Also, something else came to mind: some of the exit nodes have 100Mb/s network interfaces, rather than 1Gb/s interfaces. The cap you're seeing on the London node may be a result of that (or perhaps something else). I'll have to poke staff to see which nodes have which interfaces.

Re: The CryptoStorm Speed Test Thread

by turbz » Tue Jun 09, 2015 9:35 pm

They are MB/s, at least thats how chrome says.

Re: The CryptoStorm Speed Test Thread

by parityboy » Tue Jun 09, 2015 9:11 pm

@turbz

Can you clarify that the speeds you quote are MB/s as opposed to Mb/s? It would be good to keep things consistent for ease of readability. Thanks. :)

EDIT:
Connection: 8Mb/1Mb ADSL
Exit Node: Cantus (Frankfurt)

KTorrent 4.3.1 downloading the Linux Mint 17.1 torrent averaged 790KB/s (6.3Mb/s)

Re: The CryptoStorm Speed Test Thread

by turbz » Tue Jun 09, 2015 8:54 pm

Frankfurt seems to have difficulties connecting to peers when using public torrents, private tracker no problem other than speed being quite modest 6-7MB/s.
With Frankfurt Linhost test file 1000MB from netherlands worked quite well at 8.5-9.0MB/s

And the same file with London is steadily at 3.9MB/s for the whole download, the torrent never went past 70KB/s.

Without CS my torrent speeds were 9MB/s and the linhost test file 10MB/s.

I still remain hopeful that there's some solution to this, or at least a good explanation.
 ! Message from: parityboy
Edited throughput units for clarity.

Top