Virtual hosting in MetaTrader 4/5 platforms and data security - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
How would you feel about your hosting neighbour using a DLL to format your hard disk? Or leak all your neighbours' data files to yourself. Or shut down the computer. Or slow down the neighbours.
This is why the use of DLLs is out of the question on our hosting. If you want something specific, get your own VPS.
mmmoguschiy:
Вопрос про использование ресурсов. В ветке про виртуальный хостинг вы писали что использовать ресурсы нужно по минимуму. То есть ваш хостинг ни коим образом не ориентирован на так называемые HFT системы, способные генерировать огромное количество заявок в секунду?
Generate, who's stopping you?
It has nothing to do with resources in practical terms. Hft algorithms for the most part do not do any meaningful heavy analysis, no matter what others say or think.
Everything has been described above and before in the discussions of the service.
If you do not understand the unspoken principle of decent behavior and are aimed at destroying everything around you, as you think you are working with a machine, then there is not much you can do at any service.
Your questions about "what to load the CPU with" and "what to do with memory" show that you are not at all knowledgeable. I recommend you to look deeper into the field before you make such aggressive criticism.
What is this nonsense? Everyone works under their own virtual machine. Or do you think MQ started reinventing the wheel here too? No, of course not. The same webzilla rents a cloud, most likely. Or a separate server. And the virtualization tools are used out of the box. You can't format anything on your neighbours, and you can't break in either.
I recommend that you read the basic information on the service in the form of articles, as there are a lot of them published.
You are fundamentally wrong about the service - here is a detailed explanation: https://www.mql5.com/ru/forum/37453#comment_1161693
We build beautiful, powerful and cost-effective technical solutions, not crutches on virtual machines.
...
I'm trying to move a scalper to MT5 and I'm seeing profits cut by about half. I have users in the states etc where locking is banned. They open two accounts, no big deal, a little more manual gesture...
...
We have consistently built five (5) separate information and trading platforms from scratch, evolving expertise, watching our competitors die all the time, and have grown a huge ecosystem, maintaining and developing it. Twice made a run at creating a hybrid loc+netting system, spent time thinking about it, but each time with cries of "God forbid!" threw the idea away.
...
The decision to make MetaTrader 5 netting was extremely right. Now we have the pure MetaTrader 4 and MetaTrader 5, which can be developed further.
To reinforce: take a notebook and for two days try to write a hybrid combination of "a smart client with auto-trading - trade server - risk management / accounting - external integrations". Do not use the "I've seen someone else do it" method and think about it for at least a couple of days. If you come to the idea that "anything is possible", then look for an ambush - there is more than one.
...
I personally don't think it's right to have multiple strategies on one instrument at the same time.
It is safer to use sub-accounts for such things - with them there is more responsibility and tighter risk control due to direct separation of balances. In general, traders need to think more about safety in the first place, and not about greed and "I want to use the power of the deposit efficiently".
Control your greed rather than seeking solace in the self-deception of "a basket of strategies evens out profitability".
Generate, who's stopping you?
It has nothing to do with resources in practical terms. Hft algorithms for the most part do not do any meaningful heavy analysis, no matter what others say or think.
Everything has been described above and before in discussions of the service.
If you don't understand the unspoken principle of decent behaviour and aim to ruin everything around you because you think you're working with a machine, then there's not much you can do in any service.
Your questions about "what to load the CPU with" and "what to do with memory" show that you are not at all knowledgeable. I recommend you to get a deeper knowledge of the field before you voice your criticism in such an aggressive manner.
To the point:
All of the above I will definitely read.
I'm not aiming to kill everything around. Rather I am aimed at understanding your vision of the situation.
What area do you think I'm not in? What exactly do I need to study? Self-development is my second self - for sure and it can't hurt me either.
Z.U. Of course, I'm not at that level of understanding your system, because I'm not its developer. I'm asking you not to be too harsh and try to come down from your heights to our consumer level. I see marketing is not your strong point ;)
Is it acceptable for an EA to put a load on the CPU for 10-15 minutes every 20-60 minutes, commensurate with a 30% load on an Intel(R) Corporation Intel(R) Core(TM) i5-2500 CPU @ 3.30GHz?
I would like to be specific, what loads are acceptable for each account?
Is it acceptable for an EA to put a load on the CPU for 10-15 minutes every 20-60 minutes, commensurate with a 30% load on an Intel(R) Corporation Intel(R) Core(TM) i5-2500 CPU @ 3.30GHz?
I would like to be more specific, what loads are allowed for each account?
If on a single core, you can, in principle. But if by all cores, you will immediately get lower priority and then inevitable shutdown.
The limits are configurable and some of them will be shown directly on the performance graphs in red line in the upcoming builds. We are upgrading our virtual server network every day and increasing the functionality.
By the way, MetaQuotes ID limit overrun notifications from virtual servers have already started working.
If on a single core, in principle it is possible. But if on all cores, there will be an immediate decrease in priority, and then an imminent shutdown.
Limits are configurable and some of them will be shown directly on performance graphs in red line in upcoming builds. We are upgrading our virtual server network every day and increasing the functionality.
By the way, MetaQuotes ID limit overrun notifications from virtual servers have already started working.
Yes, the load on a single core is exactly the load I have on my home machine right now. Therefore the interest is purely practical. After all, the very song is to use specialized hosting for systems with auto-optimization.
No, not the song.
Dedicated hosting is for trading, not for eating up shared resources.