A disaster which brings down a company’s online operations and data records can kill the business. Losing business records can make it impossible to reconstruct obligations and receivables. Being without a website for days drives customers away in vast numbers. Fires, floods, and theft can destroy both on-premises systems and onsite backups. A disaster recovery plan is necessary to ensure business continuity.
Having a failover system in a remote location guards against the worst case. It needs to have a constantly updated record of the primary system’s data, so it can take over as soon as necessary. However, a full set of dedicated failover hardware usually is prohibitively expensive. Fortunately, it isn’t necessary. A cloud backup for disaster recovery can stay available on a standby basis and jump into action only when necessary.
Setting up a recovery strategy
A disaster recovery plan is an aspect of an organization’s business continuity plan. Setting it up requires an analysis to identify the systems which are essential to continued operation. Overlooking any necessary systems could keep failover servers from working properly.
Systems can fail in several ways when they’re most needed. A physical disaster, such as a fire, could make the machines in a data center unusable. Malware or a system break-in could destroy essential data. A hardware failure could wipe out file systems. A data center’s power supply can fail, though backup power supplies will keep it going through short outages. Its Internet service could stop operating. Larger data centers use more than one Internet service against this possibility, but a severe natural disaster could break all connections.
Assess the risk
A risk assessment is the first step. Agile IT can help you to identify the services that most urgently need protection from outages and data loss.
The ideal is complete redundancy and failover with no data loss or downtime, but it may not be the best choice when weighing costs and benefits. In most cases, it’s necessary to identify the critical systems and determine an acceptable level of an outage. Two metrics which are central to this are the Recovery Point Objective (RPO) and Recovery Time Objective (RTO).
Recovery Point Objective
The RPO is the maximum time interval over which loss of data is acceptable. If the RPO is one hour, then it’s considered acceptable to restore systems to a state which is out of date by no more than an hour. Any amount of data loss is painful, but “acceptable” means that the cost of a shorter RPO is more than the organization can justify.
Recovery Time Objective
The RTO is the maximum amount of downtime which is acceptable. An RTO of ten minutes means that a usable level of service should be in place no more than ten minutes after a system failure.
There’s no way to guarantee these objectives for all cases. A disaster could be widespread enough to affect failover systems as well as the primary ones. The objectives take reasonably foreseeable situations into account, not apocalyptic scenarios.
Disaster recovery with Azure
Businesses of all sizes can keep their operations safe using Azure disaster recovery. Several of its services work together to provide cloud backup and continued operation with minimal downtime.
Azure Backup keeps a secure copy of all data. Each server backs up to a Recovery Services Vault. The system administrator can back up the full system state and set a schedule for updating the copy. The backup can act as a virtual machine, ready to take up the burden if the on-premises systems fail. Hardware systems and VMs in the data center can both be replicated to Azure.
Set up VMware virtual servers
The best way to make full use of Azure recovery is to configure the on-premises systems as VMware virtual servers. An entire Windows or Linux VM can be replicated to Azure. When the primary VM becomes unavailable, Azure loads and launches the failover copy. It uses the same public IP address as the primary, so the change is invisible to users. The downtime is barely noticeable. Multiple VMs can be set up in different regions, further guaranteeing availability. Until there’s a need for failover, you pay only for backup, not for the cloud VM.
Failover can be partial or complete. If just one server fails, it’s possible to activate just its counterpart on Azure. It will continue to work with the on-premises VMs. It may be simpler or more efficient, though, to launch the entire backup as a unit, even if some of the hardware is still operational.
It’s also possible to back up just selected applications to Azure. This can simplify the operation, reducing the amount of backup and processing power needed to keep going. Azure site recovery includes application-aware features for some Windows applications that let them continue in a consistent state.
The DNS infrastructure has to be set up so it will switch to the cloud servers if the primary ones become unresponsive. One way to do this is to put the DNS controller on Azure.
Moving services to Azure
The option of recovering failed systems on a cloud service suggests an obvious question. Why not run essential services on Azure all the time? This is a viable alternative for many businesses and can reduce their requirements for a data center. Some companies have policies requiring them to keep their services on the premises. But if cloud backup and failover are allowable, it’s not a huge step toward running all the time on the cloud.
Using a cloud database has special advantages. It’s available whether the applications are running in the data center or on Azure. If the on-premises systems fail, the database is unaffected and will run without interruption or data loss. Keeping all essential information on a cloud database can make a zero RPO possible.
A properly configured and maintained cloud server is as secure as all but the very tightest on-premises data centers. Databases can be encrypted, with the keys only in the hands of the administrators. Even the cloud hosting company can’t acquire the data without those keys. Enterprise-level security is available for the most critical scenarios.
Cloud services with Agile IT
Whether you’re contemplating a full migration to the cloud, a disaster recovery program, or just augmenting your systems with a few cloud services, Agile IT is ready to help you plan your course. We’ll provide a rapid assessment of your infrastructure, so you can determine the most cost-effective way to use Azure services.
It isn’t necessary to do everything at once. An incremental migration plan will provide growth as needed without excessive disruption of existing operations. Whether you put services on the cloud full-time or can switch them over when needed, you get increased reliability and reduced downtime and data loss. Contact us to set up an assessment.