We understand that without a solid foundation, it is impossible to run financial-grade services or applications. This is why the entire infrastructure has been built around the combination of best-of-breed technology, and world-class data centre facilities. Some of the equipment that is used could be seen as overkill, but as performance and security are of critical importance, this is seen as a necessity.
With over 5 buildings on-campus covering over 34,000 square meters of colocation space and interconnect capability with over 400 companies, Equinix has rapidly become the industry standard for data centre hosting of financial applications; as such it was a logical location for our servers.
We have a large number of racks hosted in this facility with on-site spares and remote-hands to complete any maintenance tasks required.
Power is supplied by 2 diverse 40 Megawatt feeds from the National Grid with N + 2 diesel generators as backup.
Cooling is run in an N+2 redundancy manor to ensure both temperature and humidity in technical space is maintained to industry accepted standards and efficiency.
Fire suppression is in place across all areas with double-interlocked dry pipe pre-action.
Security is operated to a high standard with 24 hour manned security and intruder alarms in place, along with biometric identification required to access the data halls.
Without reliable connections to the Internet, the best equipment in the world is of no use. This is why we have invested heavily in ensuring that the network is always up.
For our primary IP Transit, we have partnered with Hibernia Networks who own one of the largest financial grade networks in the world. Through their networks, we have the quickest possible route to New York, and low-latency access to the key trading venues throughout the world. This overall results in faster access to the servers for clients, and quicker trade execution.
Due to the carrier neutral element of the facilities, network-to-network interconnects can be created to partners’ systems. This is one way that a reduction in platform latency can be achieved by ensuring that traffic goes direct rather than round the Internet. Whilst this is an effective approach, the preferred strategy is for 4X Solutions to host Metatrader Data centres locally. This provides zero latency for clients accessing the server, and reduces the overall bandwidth going in and out of each network. For the ultimate in connectivity, both approaches can be combined to provide a world-class service offering.
We currently have direct interconnects with a number of brokers who are marketing the service as an “Ultra Low Latency Trading Environment”. We are seeing a lot of customers wanting to fine-tune their trading in this way, as this is an approach taken by institutions for years.
We have standardised on Dell servers and Cisco LAN switches, all of which is covered by a vendor-backed 4 hour on-site response, with critical spares on-site for installation by the remote hands service.
VPS hosts are all of an identical specification to ensure a consistent user experience, regardless of which host they are running on.
- 2 x 6 Core Intel 2.8Ghz Processors with 12MB Cache
- 96GB 133Mhz RAM
- SSD Drives
- Battery Backed RAID Controller
- Dual Power Supplies
- 4 x 10GB Network interfaces (LACP Bonded)
Hypervisor & Guest OS
We have used a number of different Hypervisors in the past, but made the decision to use Hyper-V three years ago. In the past open source hypervisors such as KVM or XEN were used, which worked well, but sadly lacked a commercial support programme. Also, following some testing that was completed, it was demonstrated that Hyper-V was the best choice for hosting Windows VM’s. This is largely because the development teams have privileged access to the Windows OS developers.
The hypervisors are running Windows 2012 R2 Datacentre (64 bit) and managed by System Centre 2012 R2. Windows 2008 R2 (64 bit) Datacentre is currently deployed for guest OS’s.
Hyper-V 2012 has brought a tremendous set of capabilities including the ability to Live-Migrate VM’s; this means that if load balancing between hosts is required, or maintenance on one host, all of the VM’s can be migrated to another host live, without any interruption. Traditionally, this requires a SAN, but Hyper-V allows this to happen with local storage, which fits the architecture perfectly.
When dealing with anything of a financial nature, security is of upmost importance. This is why we protect the networks by Checkpoint firewall clusters. Checkpoint is by far the market leader in the security field, boasting 100% coverage of the Fortune 100, and 90% of the Fortune 500. A combination of traditional firewalling, Intrusion Prevention, and other techniques are used to mitigate the chance of an attack.
Security has been taken seriously from day one, which is why there has not been a single instance of the networks being compromised. Due to the sensitive credit card data that is stored for billing purposes, we are also regularly PCI DSS audited / scanned with the results reported back to the card issuers.