Business Continuity & Disaster Recovery
The Business Continuity Plan (BCP) is an essential part of any organisation’s response planning. It sets out how the business will operate following an incident and how it expects to return to ‘business as usual’ in the quickest possible time afterwards.
A BCP need not be specific to terrorist incidents and apply to any major disruption such as a major fire, flooding or power fault.
The plan itself sets out the agreed arrangements for bringing events under control, the necessary resources for maintaining critical business functions and the staff required for co-ordinating actions. It also needs to be clearly presented, avoiding vague internal references and abbreviations, and structured in such a way that people can quickly find and understand what is expected from them.
A major part of this plan is protecting your IT infrastructure. At Exponential-e we have a high-availability, IaaS platform housed within VDCs spread across two or multiple geographically diverse sites. Our IaaS platform is directly connected to our Layer 2 Core Network. The Exponential-e VDC solution supports the use of a number of technologies to facilitate high availability between the two or more sites. This means that your infrastructure could continue to work uninterrupted and without data loss, even if an entire infrastructure location became unavailable by replicating your on premise servers physical or virtual to our secure private Cloud.
The Solution Employs:
- Built-in application level replication.
- Exponential-e Server Replication.
- Layer 2, stretched subnets, spanning between Data Centres.
- File level Backup and restore is available if required to mitigate against deleted and corrupted files.
Exponential-e’s VDCs are connected to the Exponential-e Core Network. which is able to provide a low latency (typically below 2ms), high bandwidth link between sites.
This provides the ideal environment for our customers to use any existing application level replication technologies built into their critical applications, between the two or more physical locations.
We use our Server Replication service for those servers or applications that cannot be protected by in-built application level replication, or for those that our customers wish to protect using our replication technology.
Protecting mission-critical applications is not only about copying data. Rather, it is about ensuring application recovery, protecting from logical errors and frequently testing failover to ensure an organisation can recover the applications, settings and files when needed.
Exponential-e uses the latest server replication technologies for VMware that allow replication at the Hypervisor level. Hardware agnostic virtual replication appliances (VRA) deployed directly inside the virtual infrastructure continuously replicate data from selected virtual machines by capturing the IO stream. Therefore, each time a machine writes to its virtual disks, the write command is captured, cloned and then sent to the recovery site. This gives you journal based, any point in time recovery, without the need for continuous snapshots. All of this is managed at the vCentre level and allows for one click failover and failback of one or multiple servers.
Server Replication is offered as a managed replication service that is designed to run across Layer 2 Networks to enhance the predictability of Recovery Point Objective (RPO). Using this solution has extensive benefits, these are outlined below:
- No need to change IP addresses: due to operating the solution over Layer 2 the IP addresses can remain the same which means there is no need to change DNS records on failover.
- Bring online one or multiple servers: the Business Continuity solution runs as part of your own Network, therefore you have the ability to restore an individual server all the way through to restoring an entire site the VMware Vapps are protecting.
- RPO and Recovery Time Objective (RTO) tailored to applications: you can specify the RPO and RTO for your applications so that your business continuity plan fits business requirements.
- One click failover and failback: Allowing for one click recovery to the business continuity site and one click failback to the primary site once normal service is resumed. All with minimal disruption to your end users due to the advantage of operating over Layer 2.
- Protect entire applications by selecting the individual VMs that are part of the application.
- Consistent application protection and recovery with our Virtual Protection Group, with or without volume shadow service (VSS) support.
- Test failover with the click of a button, while maintaining protection of your production application.
- Recover from logical errors with ‘point-in-time’ recovery, enabling you to failover to a previous point in time up to five days.
- An RPO measured in terms of seconds, not minutes, with the use of continuous replication.
- Recover applications instantly and automatically, providing an RTO of minutes, not hours.
- Protect thousands of virtual machines with scale-out replication architecture.
- Reduce application performance impact to zero by eliminating snapshots.
- Recover from logical or operator errors with journal-based, point-in-time recovery.
- Monitor and report on the status of your server replication, with built-in reporting and performance user interface.
- Support for VMotion and Storage VMotion.
- Transmission, compression and bandwidth scheduling features ensure ideal performance over the WAN.
- vCloud Director integration, seamlessly protects and recovers vApps.
- Automation: failover, failback and testing are all automated, both in vCenter and vCloud environments.*
The centralised management of the replication solution simplifies tasks, ensuring application availability is optimised. Server replication supports up to five days’ worth of multiple recovery points, depending on the change rate of the server. This allows recovery from corruption of the primary servers’ data, applications and operating system. The solution is illustrated below:
Exponential-e’s solution allows our customers to meet an RTO of four hours with an RPO of any point in time over the previous five days. Exponential-e believes that the Server Replication service offers the best ROI for our customers to meet its target RPO and RTO. From a hardware perspective we use enterprise grade Servers with multiple power and Network feeds and distribute our IaaS environments across a number of blades to ensure no single hardware or network failure can disrupt your service. Need to understand more about RTO & RPO? Click here.
Testing of Server Replication
Both Server Replication and the vCloud Director Orchestration layer enables our customers to start and provision existing VMs in isolated VLANs to test server replicas, disaster recovery, and testing for critical systems in an independent, isolated environment away from the production VMs.
Experience and Customer Examples
Of the customers that use our VDC environment on an IaaS basis for dual Data Centre, critical application hosting, the majority take advantage of our Server Replication service for applications which are not equipped for in-built replication/high availability. Customers that rely on an SQL server for their back-end database frequently configure their solutions for clustering or mirroring, leveraging our low latency Network to create a stretched VLAN between Data Centres. We also have examples of customers that have chosen to configure their Intranet front-end web layer as a (virtual) server farm, shared across two of our Data Centre sites, with a low-latency stretched VLAN in between. The server farm appears to the end users as if all servers are on the same subnet.
The capabilities of the Exponential-e Network and VDC Infrastructure, combined with the use of application-level, server-level replication and online-backup will provide you with a safe, secure, high performance environment for hosting critical applications. The proposition described above presents just one of a number of possible scenarios that can be refined at a subsequent design discussion stage of the process.
*Subject to configuration.top