Last week cTrader released another new cAlgo which fixed a bunch of features which our copier required. We have now released version Beta 1.1 – this has a few nice core changes (nothing that will make any difference to the trading) so please give this a test if you are part of our testing program. You can download the beta from the downloads area – this will only work on demo platforms.
For the last few weeks we have had to close new registrations and subscriptions to our network due to a huge increase in demand. Over the last 10 days our team has been working around the clock to come up with a new solution which puts a no limits infrastructure in place. We have now successfully implemented a new backend infrastructure which utilizes multiple redundant load balancers to successfully spread load across each of our DB clusters.
We have also change the database cluster and we are seeing results which are 10x better than previous.
What this means to you?
We will open up registrations again and anyone who was seeing slower execution times should now be back to normal.
We have been working on a new EA which utilizes a new connection method. This will hopefully reduce latency by a further few milliseconds when connecting to our EA servers. In real terms, we won’t see anything, but every millisecond we can save will help!
We are doing some major changes to the way that our EA infrastructure works and as such one of these changes now means that the heartbeat changes from showing sub 5 seconds to showing sub 30 seconds.
The heartbeat is only a visual aid to help you know that your EA is online and does not in any way reflect the speed at which the copier is checking for trades. We have changed it to heartbeat less often so that we are not having to write as much data to our master databases.
Unfortunately the server upgrades have not reduced the current issues with our system.
The problem is incredibly perplexing. To give you all a brief run down of the situation:
- Mid May our server were running perfectly with plenty of spare capacity.
- 2 Weeks ago we obtained a new client that attached a significant number of connections which pushed the server load to capacity.
- We have installed new, extremely robust hardware that has reduced our load to around 30% (leaving us with a theoretical 70% spare capacity)
- The issues we were having as a result of the overload have not been resolved.
So we’re our servers are running at 30%, however they’re behaving as though they are running at capacity…. here in lies our confusion.
We have contracted a MySQL expert (at great expense) to help us investigate this conundrum.
Our team have been working around the clock to find a solution to this problem and continue to do so.
Unfortunately we are not able to commit to a specific time frame for which the problem will be resolved and allow new subscribers to join the network.
In the meantime new subscriptions remain frozen.
It’s important to note that existing subscribers should not experience any issues with their current connections as a result of this problem.
Apologies for the delay, however we are still configuring our new hardware.
It’s a much more difficult task than we anticipated. After working all weekend there are still a number of tasks to complete to make sure the network is rock solid when we open the doors again.
New subscriptions will resume on 3rd of June.