Technology

Cloud Dialogs business continuity plans are based on two core principles;

  • Redundancy and elimination of single points of failure from the network
  • Robust backups and recovery procedures

Our main emphasis is on the first of these;

Core components of the Cloud Dialogs Architecture

World class data centre – Cloud Dialogs network runs on dedicated hosting class hardware in a private Cloud within a Tier 1 fully managed London datacentre.

Redundant Hardware – Within our private cloud we have multiple separate physical servers which are hosting class (i.e. the underlying hardware used by datacentres if you buy a dedicated Virtual Server).

Our solution then runs on a variable number of Virtual Machines which are distributed across these physical servers. The servers run Hypervisors which actively and continually monitor the physical servers, in the event of a complete hardware failure the VMs that were on the machine in question are automatically spun up one of the other physical servers. Each server is very easily capable of running our entire network and every customer with at least 2 years envisaged growth potential.

Each physical server runs RAID10 so that disk failure does not cause any interruption to the service, in the event of a disk failing it is automatically detected and swapped out

The servers also have multiple redundant power supplies to eliminate that area of failure.

SQL Server SAN storage – All databases reside in a separate SAN that is automatically distributed and replicated across multiple servers as a core service component.

Regular backups – Customer databases are backed up overnight on a daily basis to a separate storage facility and log shipping runs on a transaction by transaction basis to enable us to recover up to the most recent transaction if needed. In our six years of operation to date we have never had to resort to rebuilding from this backup however we can completely rebuild a customer environment within 2 hours should the need arise. The backups are monitored and automated reporting confirms successful completion. We test recovery from this backup source once a month.

Restricted access – We do not provide direct access to the databases and each customer runs on a separate SQL Server database. Our build processes are all fully automated and we never live patch the production system, updates are only ever performed through the full upgrade process.

Pre-production environment – All changes are deployed to an internal UAT environment for testing and then to a pre-production environment for final testing. The pre-production environment is identical to the main production platform and could run all live customers. We never upgrade the pre-production environment and the production environment at the same time and we have the ability to automatically move accounts between these platforms. We have not ever had to utilise this but in the event of a critical failure on the production platform we can move customers to this backup platform to restore system availability. We routinely move internal accounts between platforms to ensure this is all operating normally.

Virtual office – All internal Cloud Dialogs data and systems are Cloud based. All source code is automatically and continually stored off-site using secure source code tracking systems. We can operate every aspect of the business from any location with an Internet connection so there is no risk from physical loss of computers, disks or data from any physical office.