CNS VPS service

[CanadianPsycho]

I figured maybe we could use a thread to centralize discussion about Commercial Network Services' VPS service.  I've noticed that Barry is as prolific and helpful here as he is on other forums so maybe it would be favourable for both him and us as a group to have a spot where questions can be asked and hopefully answered.

I have one question on my mind at the moment that hopefully Barry could help me out with.

I don't know what kind of latency I'm getting to the GoMarkets servers in SD.  As any client of GM who uses a VPS will know, they don't give out the IP addresses of their various servers because they don't want them to fall victim to attacks as I think they have experienced in the past.  As such, I'm not sure how I can test my latency to the GoMarkets servers.  I know that it's supposed to be very low on average but I would very much like a way to test it out myself from my own VPS installation.

Is there a way to test latency from CNS San Diego to GoMarkets' SD servers?

Cheers

[BarrySDCA]

Hi CanadianPsycho,

We have GO Markets servers in our San Diego and UK datacenters.  So latency to these servers from your CNS VPS will be <1 ms.

IP addresses are listed here:

http://helpdesk.commercialnetworkservices.net/index.php?_m=knowledgebase&_a=viewarticle&kbarticleid=174

The SDCA server is listed in your MT4 terminal as DC3.
The UK server is listed in your MT4 terminal as DC5

From what I understand, these servers talk to their main trading server.  I can completely understand why they would not want to give out their main trading server IP, but the MT4 terminals have got to talk to someone.  Ideally, nobody should be talking directly to the main trading server anyway (for security reasons).  I am not 100% in-the-know on the GO Markets internal configuration.  I know they have two main servers, with one being a backup.

[CanadianPsycho]

Thanks Barry.  I totally missed the IP addresses on that chart.

Cheers

[robl45]

CNS looks like a good choice, but the prices are crazy high,  I had to take a nitro pill to prevent going to the emergency room after reading the prices.

If my thrustvps starts to have issues,  I'll look to switching, but for now I can't justify going from less than 8 dollars a month to 30+ dollars a month for the same thing.

[BarrySDCA]

CNS VPS and Thrust VPS are *far* from the same thing.

[robl45]

CNS VPS and Thrust VPS are *far* from the same thing.

could you elaborate on that?

[donm]

CNS has lost it's MT4 email alert connection.   Contacted service and they mentioned the email is backed up with a "flood" of emails.

Is there an alternate way to receive emails?

Otherwise CNS has been first rate and service is excellent.

donm

[BarrySDCA]

robl:  I saw your message yesterday and signed up for service so I can give you a more complete response.  I already have a laundry list to detail back to you.  Unfortunately, I am back in the office this morning and after 12 hours of waiting for my vps to provision with them, it's already experiencing issues and so now I must contact them by support and straighten it out.  I will follow-up on your excellent question as soon as I am actually able to login to their vps.

donm:  Our email gateway has been flooded 2 times in the past month or so.  I believe it must be a new EA or indicator that is sending these.   and I'm talking tens of thousands in a matter of minutes.  first we doubled capacity and last night we doubled it again, but our techs are still clearing it out - as of this writing ~26,000 more messages to go (a fraction of what was there).

If you contact CNS support, we can help you to fixup your DNS and relay out your own VM while we figure out a way to curtail this abuse.   Otherwise, I expect the mail delivery time to be back to normal next 1 hour or so.

One thing to keep in mind:  if your terminal is sending tens of thousands of emails to one address in a short period of time, then your email provider will eventually delay or disable our ability to relay to you.  In this case, these messages were going to hotmail, who started to delay our message delivery to them.  we have already worked around it by setting up more relays.  next stop, curtail this abuse.  rest assured we are working dillegently on it.

[donm]

robl:  I saw your message yesterday and signed up for service so I can give you a more complete response.  I already have a laundry list to detail back to you.  Unfortunately, I am back in the office this morning and after 12 hours of waiting for my vps to provision with them, it's already experiencing issues and so now I must contact them by support and straighten it out.  I will follow-up on your excellent question as soon as I am actually able to login to their vps.

donm:  Our email gateway has been flooded 2 times in the past month or so.  I believe it must be a new EA or indicator that is sending these.   and I'm talking tens of thousands in a matter of minutes.  first we doubled capacity and last night we doubled it again, but our techs are still clearing it out - as of this writing ~26,000 more messages to go (a fraction of what was there).

If you contact CNS support, we can help you to fixup your DNS and relay out your own VM while we figure out a way to curtail this abuse.   Otherwise, I expect the mail delivery time to be back to normal next 1 hour or so.

One thing to keep in mind:  if your terminal is sending tens of thousands of emails to one address in a short period of time, then your email provider will eventually delay or disable our ability to relay to you.  In this case, these messages were going to hotmail, who started to delay our message delivery to them.  we have already worked around it by setting up more relays.  next stop, curtail this abuse.  rest assured we are working dillegently on it.

OK,  email alerts starting from last night, starting to arrive.  These are from demo account,  I can wait for you to clear the backup.

donm

[BarrySDCA]

it is down to less than 1000 to go.  I am keeping a close eye on it.  Next time you need help, please do not hesitate to open a support ticket.  Our support response is crazy fast.

robl:  I opened an emergency ticket with thrust over 1/2 hour ago.  still no reply.  I will advise after I have a better comparison for you.  really this is a very easy question but I am sure there are more things in which they lack that I do not know about, and so I would like to give you a more complete response.

[e1vis]

Barry

Are there any plans to set up an Australian (or Asian) datacentre? I am with an Australian broker and, whilst I really don't want to consider leaving CNS (or changing my broker) the latency is very high at the moment.

cheers

[BarrySDCA]

We explore the subject every few months.  As of this moment, we have no plans to open an Australian DC.  We routinely invite Australian brokers to host a server with us, and to date only GO Markets has done that with a server in our San Diego and UK datacenters.

Brokers should ideally be close to their liquidity providers.  What is the point of having very low latency to a broker if, on their back end, they are traveling 1/2 way around the world to a major financial center in order to reach their liquidity partner and execute your trade?  This is why many brokers do not have their servers in the same location as the business.  One example off the top of my head:  FX SOL was an Australian broker up till few months ago.  Now they are UK.  Latency was and has always been <2ms from our NYC datacenter, which is just 2 blocks off wall street and close to their liquidity partners too I'm sure.

The issue of Australia is that on the network map, it is extremely isolated.  The entire continent has only two feeds to the rest of the world.  One to Asia and the other to the western USA.  And there is the problem - bandwidth is extremely expensive - over 600% what we pay in the USA.  Additionally, due to the latency issue, we do not see Australia really taking off with hosted MT4 servers.  The brokers in Australia are relatively young, and popular with Australians.  as they grow globally, they will realize that they need their servers elsewhere, which would leave us holding the bag for a huge datacenter investment and nothing to put in it.

Getting GO Markets to our DC's was largely an effort of our mutual subscribers.  Several months ago Pepperstone signed up for our broker bundling program, so they could sponsor trader's vps subscriptions, but I do not believe they have done anything with it.

[e1vis]

We explore the subject every few months.  As of this moment, we have no plans to open an Australian DC.  We routinely invite Australian brokers to host a server with us, and to date only GO Markets has done that with a server in our San Diego and UK datacenters.

Brokers should ideally be close to their liquidity providers.  What is the point of having very low latency to a broker if, on their back end, they are traveling 1/2 way around the world to a major financial center in order to reach their liquidity partner and execute your trade?  This is why many brokers do not have their servers in the same location as the business.  One example off the top of my head:  FX SOL was an Australian broker up till few months ago.  Now they are UK.  Latency was and has always been <2ms from our NYC datacenter, which is just 2 blocks off wall street and close to their liquidity partners too I'm sure.

The issue of Australia is that on the network map, it is extremely isolated.  The entire continent has only two feeds to the rest of the world.  One to Asia and the other to the western USA.  And there is the problem - bandwidth is extremely expensive - over 600% what we pay in the USA.  Additionally, due to the latency issue, we do not see Australia really taking off with hosted MT4 servers.  The brokers in Australia are relatively young, and popular with Australians.  as they grow globally, they will realize that they need their servers elsewhere, which would leave us holding the bag for a huge datacenter investment and nothing to put in it.

Getting GO Markets to our DC's was largely an effort of our mutual subscribers.  Several months ago Pepperstone signed up for our broker bundling program, so they could sponsor trader's vps subscriptions, but I do not believe they have done anything with it.

Many thanks Barry - I'm with Pepperstone so hopefully they will address this issue.

[LFN]

We routinely invite Australian brokers to host a server with us, and to date only GO Markets has done that with a server in our San Diego and UK datacenters.

Did you invite FinFX to host a server at your UK datacentre?
Seems odd that FinFX main server is in Finland and there is not possible getting better latency from a VPS than about 45ms.

[BarrySDCA]

We did invite FinFX to host a server in our UK datacenter, just a few week ago.  They were very keen to the idea but concerned about cost.  We offered to GIVE them a server to use, power and licensing, and only charge them actual cost for bandwidth, which IMHO was not all that much.  Even less actually because bandwidth to CNS subscribers is not billed.  They declined due to the cost.

We also offered to host their main trading server in our UK datacenter, but they said the law in Finland did not allow their main server out of the country.

It will likely require pressure from our mutual subscribers to get them to host a server.  Still, I can't help but find it odd that they declined such a ridiculously cheap offer.