running esignal on a windows cloud server

R

Runningbear

This is not quite an automated trading question but a step on my journey to full automated. I currently use esignal for generating trade signals and I'm thinking of running the application from a windows cloud server.

This way I can ensure maximum internet uptime for rock solid signal reliability. Plus I can use esignal from my mac or ipad using remote login.

The problem I'm having is that I'm trying to calculate my likely monthly cost for such a server.

Cloud companies like Opsource charge per cpu. My i7 chip in my home computer spikes to 15% every couple of minutes. The rest of the time its at zero. If my i7 chip has 4 cores, what would this convert to in terms of cpu requirement for a cloud based server? Is this still 1 cpu? Or does a 4 core cpu convert to 4 cpu's in a cloud set up?

Thanks in advance for any help.

Runningbear
 
VMs and Bare Metal are truly different animals.

If you are going automated you can either get free VM trials or you should know what the system requirements are. They perform just like a desktop in terms of used/free resources but they also can run at optimized speeds.

You can't really compare them.

I host VMs for a couple hedge funds. If you want you are more than welcome to demo a dual core vs. a quad core VM running VMware ESXi 5.0. You are more than welcome to do benchmarks and other tests to just get a feel for what the VMs do in terms of horsepower vs. a desktop. (virtual vs. bare metal)

If you say yes to the offer I will say I'd feel most comfortable if you didn't put anything sensitive or proprietary on there.
 
...because they are demo or is this your general business policy?

Do you have a link to get some info on pricing for your services?

Thanks


Quote from WinstonTJ:

VMs and Bare Metal are truly different animals.

If you are going automated you can either get free VM trials or you should know what the system requirements are. They perform just like a desktop in terms of used/free resources but they also can run at optimized speeds.

You can't really compare them.

I host VMs for a couple hedge funds. If you want you are more than welcome to demo a dual core vs. a quad core VM running VMware ESXi 5.0. You are more than welcome to do benchmarks and other tests to just get a feel for what the VMs do in terms of horsepower vs. a desktop. (virtual vs. bare metal)

If you say yes to the offer I will say I'd feel most comfortable if you didn't put anything sensitive or proprietary on there.
More...
 
Quote from amazingIndustry:

...because they are demo or is this your general business policy?

Do you have a link to get some info on pricing for your services?

Thanks
More...
I don't know what your issue is with me but following me around and posting aggrevating comments is getting really old.

You clearly don't know much about virtualization if you are asking or implying that a VM is only slower because it's in a demo environment.

To answer your question, bare metal will always be faster. The hypervisor takes up some system resources as well as introduces some latency. If you were to benchmark two identical systems, one bare metal and one with any flavor hypervisor (vmware, citrix, Xen) the bare metal box would always be slightly faster.

Regarding proprietary info installed to a demo VM, they are equally as safe as production. I just say that so that #1 there is no confusion going forward about who's property is who's and #2 if you are getting a demo it should be for testing compatibility and connectivity, not for programming & development.
 
Quote from ProTrader432:

Give Windows Azure a try.

http://www.windowsazure.com/

The Advantage is that you can test it out three months for free

I use it for my Algo Servers and it's great
More...

Protrader, can you tell me if the memory allocation includes the memory required to run windows server? Or if this separate?

thanks,

Runningbear
 
With all due respect but could it be that you are easily agitated? This is now the second or third post in a few days where you try to read behind the lines and no matter what try to dig up dirt but actually end up completely misunderstanding the points made.

I was referring to this line of your earlier post: "If you say yes to the offer I will say I'd feel most comfortable if you didn't put anything sensitive or proprietary on there."

Thats why I asked whether this is just during demo times or generally the case and yes I admit it was meant to be a bit ironic because if you cannot even safeguard against IP theft when you let clients demo the product why should they have confidence their properties are protected when they run on paid services.

Winston, you are hugely contradicting yourself. You say on one hand that stuff is absolutely safe and people who spread IP theft "rumors" are liars then you tell us above and in your latest post that " I just say that so that #1 there is no confusion going forward about who's property is who's ". So if its safe and protected from prying eyes then why could there ever be any confusion??? Seriously dude, I am not attacking you nor am I following you around but you kind of open yourself up to a lot of raised eyebrows with your very weird comments.

And yes, I do not know much about virtualization nor have I ever claimed I do, nor have I ever made a reference to any issue regarding virtualization.

You were unclear about IP before, then you are unclear about IP when you talked about your own services you offer, including demos, and you were unclear when you attack someone who complained his IP was stolen. Maybe you wanna try to be crystal clear then there won't be anyone second guessing you. Right now you come across as a very opaque person who shies away from making things clear.

Quote from WinstonTJ:

I don't know what your issue is with me but following me around and posting aggrevating comments is getting really old.

You clearly don't know much about virtualization if you are asking or implying that a VM is only slower because it's in a demo environment.

To answer your question, bare metal will always be faster. The hypervisor takes up some system resources as well as introduces some latency. If you were to benchmark two identical systems, one bare metal and one with any flavor hypervisor (vmware, citrix, Xen) the bare metal box would always be slightly faster.

Regarding proprietary info installed to a demo VM, they are equally as safe as production. I just say that so that #1 there is no confusion going forward about who's property is who's and #2 if you are getting a demo it should be for testing compatibility and connectivity, not for programming & development.
More...
 
Running off the cloud is nice. When I was looking at Amazon AWS (and may be true for other cloud services) keep in mind you either pay to run it 24/7, or shut it down when not in use, and pay for storage to persist the system image. Running 24/7 is more convenient.

If you can verify in task manager that it is the esignal process that sometimes goes to 15% CPU on your quad-core, then worst case, that would translate to 60% on a single CPU or 30% on a dual core.
 
Actually look at task manager performance tab, it probably shows 8 cores (due to hyperthreading). If so then worst case, your 15% = 120% on a single core. In other words not fast enough.
 
You must log in or register to reply here.
Back
Top