Disaster recovery is currently mostly carried out by replicating applications between primary and secondary data centers. When the primary data center experiences a disruption that interferes with the organization’s business continuity, the secondary data center will take over the functions of the primary data center and activate the replicated applications. Then, with virtualization, organizations can then perform replication at the virtual machine level. One of the advantages of virtualization in a disaster recovery scheme is that the secondary data center can be passive (virtual machine), inactive, but the organization can still get adequate RTO.
Because secondary data centers are passive, Disaster Recovery services are highly sought after by service providers and organizations that want this service. From the service provider’s perspective, they do not need to provide 1:1 resources (unless they have to follow certain requirements from service users) because the virtual machines in the data center are passive (in a powered-off state). Meanwhile, from the perspective of organizations that want this service, they will certainly get very efficient costs because service providers should be able to implement resource sharing methods.
There are several important things to consider when choosing a Disaster Recovery service, namely:
RPO (Recovery Point Objective)
This regulates how long data is allowed to be lost when a disaster occurs. An RPO of 15 minutes means that the system will replicate data from the primary data center to the secondary data center every 15 minutes. RPO is influenced by two things, namely: the amount of data change (data delta) and the amount of bandwidth for replication. As a reference, with a 1 Mbps (megabits per second) network, we can transfer 300 MB (megabytes) of data in 1 hour (60 minutes). Therefore, to achieve an RPO of 15 minutes, an organization needs approximately 4 Mbps for 300 MB of delta data. To achieve an RPO of 0, a synchronous replication solution is required.
RTO (Recovery Time Objective)
This regulates how long the secondary data center can operate after a disaster occurs at the primary data center. RTO will be affected by the time needed to determine the disaster conditions, coordination with DR service providers, initialization of the secondary data center (application initialization), and network reconfiguration from the primary to the secondary data center. Considering the application and backup techniques, some manual steps such as application verification may still be required.
Performa Pada Disaster Recovery
In order for DR services to be utilized effectively, these services should have a minimal impact on every application located in the primary data center. DR service providers must ensure that the software they use to protect applications located in the primary data center has a minimal impact. In addition, service providers are also responsible for performance when the secondary data center of DR service users is activated. DR service users should conduct DR test scenarios after using this service to align the expectations of service providers with those of DR service users.
Consistency
DR service providers must be able to guarantee data consistency when it is used (in the event of a disaster). This data consistency includes application consistency, transaction consistency, and point-in-time consistency (consistency at a specific point in time). On Windows OS, this consistency is usually supported by Volume Shadow Copy Services (VSS).
Location in Disaster Recovery
The location of the secondary data center is also an important factor in selecting a DR service. Government policy on data storage stipulates that the distance between the primary and secondary data centers must be at least 30 km. Distance also affects WAN costs and increases latency. With regard to latency, users usually opt for asynchronous replication. To achieve synchronous replication over such a distance, additional costs and equipment are required to reduce latency.
Finally, DR service users must carefully assess the readiness of DR service providers. This readiness includes the DR service provider’s contact points, namely 24×7 helpdesk or technical support, DR preparation schemes (DR team readiness, DR service provider team structure, and DRP), and ensuring that service providers can facilitate DR trial scenarios to ensure a smooth DR process in the event of a disaster.
Best regards!
Garibaldi – Senior Cloud Consultant IndonesianCloud
That concludes our explanation. If you would like to read other articles about technology or want more information about Indonesian Cloud products, you can visit our website Indonesiancloud.com and our VPS website cloudhostingaja.com. See you in our next article.