New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Looking for KVM: 1024-2048MB RAM | 2GB SSD | 20GB HDD | h/w RAID10 w/BBU | 500GB b/w | 1xIPv4
This is for 2 small DB clusters (<= 1 million records). Need 6 of these to start with, different locations or DCs.
Preferred locations:
(please provide test IP and also looking glass, if available)
US: Piscataway, Philadelphia, Atlanta, Chicago or Miami
- at least one with <80ms latency to EU
- plus some place with very good latency to Hong Kong/SE Asia - <200ms?
EU: Germany, Netherlands, France or UK (at least one with <80ms latency to US above)
CP: SolusVM
Budget:
1024MB RAM + 20GB HDD <= USD 7/each
1024MB RAM + 2GB SSD + 20GB HDD <= USD 9/each
2048MB RAM + 20GB HDD <= USD 10/each
2048MB RAM + 2GB SSD + 20GB HDD <= USD 12/each
Also interested in white label RESELLING - please see separate post here.
Feel free to PM me at [email protected] or SKYPE (businesscentre.hk)
Comments
< 80ms USA <-> EU? Good luck finding that one.
I've never had anything below 100.
I've never had anything below 100.
It is certainly possible to get the latency @andrzej wants, but only with a data center on the east coast.
Take for example GetKVM NJ (Looking Glass) to Avenue Host (LDEX datacenter). The ping is ~70ms.
Well, I try hard to ask only for things realistic
Try this to, say, Dusseldorf or Rotterdam or Maidenhead...
Oh. Indeed. Sorry then
I have aprox 80-90 ms to my ipxCore and serverdragon. (On a good day), usually somewhere like 90-100ms.
From sweden that is.
Inb4 rage:
Chicago VPS?
KVM?
I've never had anything below 100.
@BlueVM has 98.6 ms ping to OVH. (s1.c12.ny.bluevm.com)
Oh, and our LA location (which will soon have KVM's) have 174 ms ping to Hong Kong. (s1.c12.ca.bluevm.com)
@andrzej For the most part you should expect even with some of the best networks on the east coast to see ~70-78ms to UK, ~80-105ms to DE/NL
Choopa is gonna be the only transit provider I know that will make it to DE/NL in 80-85ms most others, including Internap/Phoenixnap in Ashburn, VA (Which usually have really nice routes) are going to be closer to 90-100ms. The main reason Choopa does better is they have a high commit with Level3 and Atrato to the EU and direct transit so they are a bit lower latency. Most others will be peering somewhere before heading to EU which is why you see the extra 10-20ms overhead.
For example, @Nick_A (Ramnode) which is in Atlanta has some pretty nice routing, however, to go to NL it peers to ntt to get to Atrato so it adds 20ms. It seems for DE they hand off to Tinet, and Tinet tends to use cheapest route first, so likely you end up with Cogent (~130ms). The throughput is still awesome though.
Also anyone who is peering Cogent to Europe, automatically expect it to be at least 100ms. This would include providers such as Colocrossing which now use a lot of Cogent preference in their mix.
There are several companies that use Choopa.net here, my suggestion is with @Fliphost for OpenVZ, @GetKVM for KVM. Also, for XEN in Choopa with 10gbps drops to the VPS servers (provides better latentcy than 100mbit or gigabit drops), try out Greeny Hosting. When I got my server with them it was on special and I got it at $7.00/month (256Mb RAM, 256Mb Swap, 15gb HD) I am not sure if they are still running that special now though.
TL;DR:
Your best bet is going to be with a provider in Choopa.net as they have direct transit which allows them to have a lower ping by ~10-20ms. Most other networks you should expect between 90-105ms to Europe from east coast.
Yes, I'm thinking of getting one VM with @GetKVM in NJ. Then something on the other end in UK with good latency to Germany and France.
Many thanks for explaining how things work
@Thelinuxbug ColoCrossing's Buffalo datacenter has excellent connectivity and low latency to Europe. Allow me to demonstrate:
Tracing route to ic.co.uk [195.62.198.11]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms 10.8.0.149
2 <1 ms 1 ms 1 ms host.colocrossing.com [198.12.64.81]
3 21 ms <1 ms 1 ms te7-4.ccr01.buf02.atlas.cogentco.com [38.122.36.45]
4 3 ms 3 ms 3 ms te0-0-0-32.ccr21.yyz02.atlas.cogentco.com [154.54.43.61]
5 10 ms 10 ms 10 ms te0-2-0-2.ccr21.ymq02.atlas.cogentco.com [154.54.29.41]
6 80 ms 80 ms 80 ms te0-4-0-4.ccr22.lpl01.atlas.cogentco.com [154.54.44.210]
7 81 ms 80 ms 81 ms te4-2.ccr01.man01.atlas.cogentco.com [154.54.36.58]
8 80 ms 81 ms 80 ms 149.6.11.26
9 81 ms 80 ms 80 ms pm3-1.keele.netcentral.co.uk [212.57.224.6]
10 89 ms 83 ms 81 ms tc-rh-csw1-g1-0-1.ic.net.uk [195.62.192.94]
11 81 ms 81 ms 81 ms reverse.keele.netcentral.co.uk [195.62.192.22]
12 81 ms 81 ms 81 ms 195.62.198.11
13 81 ms 81 ms 82 ms 195.62.198.11
Trace complete.
Tracing route to globalswitch.com [94.102.158.194]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms 10.8.0.149
2 <1 ms <1 ms <1 ms host.colocrossing.com [198.12.64.81]
3 <1 ms <1 ms <1 ms buf-b1-link.telia.net [213.248.96.41]
4 9 ms 9 ms 9 ms nyk-bb1-link.telia.net [80.91.246.37]
5 81 ms 157 ms 95 ms ldn-bb1-link.telia.net [213.155.133.148]
6 79 ms 79 ms 79 ms ldn-b1-link.telia.net [80.91.248.89]
7 81 ms 81 ms 81 ms ten2-1-0-101-t40-br3.router.uk.clara.net [213.248.75.74]
8 80 ms 80 ms 80 ms ten1-0-0-t40-cr2.router.uk.clara.net [195.8.86.141]
9 80 ms 80 ms 80 ms g6-1-t40-ar12.router.uk.clara.net [195.157.6.173]
10 81 ms 81 ms 82 ms gi5-9.ir2.str.uk.idllt.net [195.8.78.134]
11 82 ms 82 ms 82 ms vm11306.vps.uk.clara.net [94.102.158.194]
Trace complete.
Tracing route to ic.co.uk [195.62.198.11]....
13 81 ms 81 ms 82 ms 195.62.198.11
11 82 ms 82 ms 82 ms vm11306.vps.uk.clara.net [94.102.158.194]
ColoCrossing in Buffalo's connectivity and latency to Europe pale in comparison to HostVirtual in NYC (60 Hudson). Ping stats from 60 Hudson:
5 packets transmitted, 5 received, 0% packet loss, time 4005ms
rtt min/avg/max/mdev = 74.638/74.799/74.950/0.324 ms
5 packets transmitted, 5 received, 0% packet loss, time 4007ms
rtt min/avg/max/mdev = 70.331/70.597/70.804/0.171 ms
HV-NYC-60 Hudson to London:
1 69.169.80.226 0.266 ms 0.230 ms 0.214 ms
2 69.169.80.6 0.297 ms 0.299 ms 0.288 ms
3 216.221.158.249 0.470 ms 22.929 ms 0.448 ms
4 89.149.180.50 70.852 ms 141.136.110.2 71.090 ms 89.149.182.121 71.449 ms
5 77.67.95.58 70.786 ms 71.582 ms 71.240 ms
6 89.187.93.20 71.235 ms 71.415 ms 72.030 ms
7 85.13.237.202 72.597 ms 72.315 ms 71.969 ms
8 89.32.144.12 71.102 ms 70.271 ms 70.942 ms
9 89.32.144.101 71.194 ms 70.813 ms 71.126 ms
HV-NYC-60 Hudson to Amsterdam
1 69.169.80.226 0.331 ms 0.292 ms 0.282 ms
2 69.31.34.221 2.264 ms 2.257 ms 2.245 ms
3 69.31.34.127 4.679 ms 3.648 ms 3.650 ms
4 69.31.94.134 1.323 ms 69.31.94.138 1.187 ms 1.296 ms
5 204.245.38.170 76.269 ms 77.257 ms 77.301 ms
6 62.212.80.102 78.220 ms 78.165 ms 78.460 ms
7 85.17.96.203 77.091 ms 76.124 ms 77.077 ms
8 46.102.240.10 77.548 ms 76.641 ms 77.290 ms
HostVirtual isn't in the OP @andrzej 's price range but 60 Hudson does beat ColoCrossing Buffalo and Choopa when it comes to low latency to Europe
If the 10 msec is important to you, then you have to host in NYC/NJ. And we offer that too ;-).
Buffalo is a better option for 99% of users because the 10 msec isn't worth the big premium of hosting in the NYC region. We basically put financial/banking customers only in NYC these days.
+1 - 10 milliseconds isn't the only difference.
Random routing, over-saturation of switches and premium bandwidth matter.
@jbiloh Sorry, let me clarify for you, your routes to .DE and .NL usually prefer Cogent. Yes, the UK only sees about a +5-7ms latency hike to Choopa (As it doesn't ALWAYS prefer Cogent). Do that same test to Germany/Netherlands and let me know if you see less than 85-90ms, that will be the proof you need to substantiate your claim over what I said.
For example:
1 citrxx.ipxcore.com (198.144.190.x) 0.026 ms 0.035 ms 0.014 ms
2 ipxcore.aggr.buffalo.nwnx.net (75.127.1.81) 0.803 ms 1.001 ms 1.453 ms
3 buf-b1-link.telia.net (213.248.96.41) 0.674 ms te7-4.ccr01.buf02.atlas.cogentco.com (38.122.36.45) 198.047 ms 207.86.157.13 (207.86.157.13) 0.589 ms
4 216.156.0.253.ptr.us.xo.net (216.156.0.253) 23.454 ms te0-0-0-32.ccr21.yyz02.atlas.cogentco.com (154.54.43.61) 3.757 ms 216.156.0.253.ptr.us.xo.net (216.156.0.253) 23.411 ms
5 te0-1-0-6.ccr21.ymq02.atlas.cogentco.com (154.54.44.113) 10.965 ms 207.88.14.194.ptr.us.xo.net (207.88.14.194) 13.288 ms te0-4-0-5.ccr21.ymq02.atlas.cogentco.com (154.54.43.218) 11.701 ms
6 te0-2-0-6.ccr21.lpl01.atlas.cogentco.com (154.54.0.69) 80.495 ms te-10-0-0.edge1.NewYork1.level3.net (4.68.63.213) 10.446 ms ae8.edge3.Chicago3.Level3.net (4.68.127.245) 13.086 ms
7 vlan52.ebr2.Chicago2.Level3.net (4.69.138.190) 113.489 ms vlan70.csw2.NewYork1.Level3.net (4.69.155.126) 98.549 ms te0-2-0-7.mpd22.ams03.atlas.cogentco.com (154.54.37.109) 90.880 ms
8 te0-7-0-4.mag21.ams03.atlas.cogentco.com (154.54.76.138) 91.286 ms te0-7-0-0.mag21.ams03.atlas.cogentco.com (154.54.76.170) 92.963 ms te0-0-0-6.mag21.ams03.atlas.cogentco.com (154.54.76.194) 90.631 ms
9 ae-44-44.ebr2.London1.Level3.net (4.69.137.77) 98.595 ms ae-5-5.ebr2.Washington1.Level3.net (4.69.143.221) 114.417 ms 113.316 ms
10 ae-41-41.ebr2.Paris1.Level3.net (4.69.137.49) 129.687 ms ae-46-46.ebr2.Amsterdam1.Level3.net (4.69.143.73) 99.508 ms ae-41-41.ebr2.Paris1.Level3.net (4.69.137.49) 115.934 ms
11 ae-22-22.ebr1.Amsterdam1.Level3.net (4.69.143.57) 99.432 ms ae-45-45.ebr1.Frankfurt1.Level3.net (4.69.143.133) 113.937 ms xxx.xxx.xx (31.xx.xx.xx) 95.030 ms
1 citrxx.ipxcore.com (198.144.190.x) 0.020 ms 0.011 ms 0.008 ms
2 ipxcore.aggr.buffalo.nwnx.net (75.127.1.81) 2.939 ms 3.386 ms 3.753 ms
3 te7-4.ccr01.buf02.atlas.cogentco.com (38.122.36.45) 118.179 ms buf-b1-link.telia.net (213.248.96.41) 0.337 ms te7-4.ccr01.buf02.atlas.cogentco.com (38.122.36.45) 118.140 ms
4 nyk-bb1-link.telia.net (80.91.246.37) 25.651 ms 25.645 ms te0-7-0-26.ccr21.yyz02.atlas.cogentco.com (154.54.43.65) 4.755 ms
5 * * *
6 te0-4-0-6.ccr22.lpl01.atlas.cogentco.com (154.54.44.214) 80.363 ms 206.111.2.222.ptr.us.xo.net (206.111.2.222) 12.962 ms 12.930 ms
7 te0-6-0-2.ccr21.ams03.atlas.cogentco.com (130.117.48.42) 91.477 ms eth4-3.core1.nyc1.us.as5580.net (78.152.44.201) 20.383 ms 20.358 ms
8 te0-6-0-1.ccr21.ams04.atlas.cogentco.com (154.54.36.134) 90.839 ms serverius-gw.ip4.tinet.net (141.136.99.30) 99.158 ms te0-1-0-4.ccr21.ams04.atlas.cogentco.com (130.117.2.62) 91.096 ms
9 xxx.xx.xx (46.xx.xx.xx) 99.637 ms
P.S. I do not usually consider the UK to be Europe (Physically I guess you can argue it is however it doesn't participate in the the Euro, etc.)
Cheers!
Thanks, guys
10ms one way or the other will not be a problem, I think. These will be DB Cluster nodes, not that much data transfer. In fact, I'm reconsidering two clusters in favor of one, at least to begin with:
DB1 in NJ, Buffalo or even Chicago (about 95/105ms to UK/Germany);
DB2 in EU, probably Germany, somewhere with good latency all around anyway;
DB3 in L.A. or Seattle "aimed" at Asia (i.e. <150ms to Hong Kong);
Since I'm doing away with the separate cluster in Europe, probably will go for DB4 in Hong Kong, instead, for good measure. This has obvious pros, but also a downside - data will have to be written to 4 DB nodes as opposed to only three.
FYI, these 3 or 4 DB nodes (all "masters") are the backend for 7 redundant webserver pairs around the world, running the same Drupal websites that have to be up 24/7/365 come hell or high water. Most of the content will be served static from cache, but each webpage has embedded tracking/internet marketing stuff from DB and about half of the pages served will next write data back to DB. Anticipate between 30-50 such reads/writes per minute.
Frankly, never played with offloaded DB's before, not to mention DB clusters, so any comments or suggestions are very welcome