Disaster Recovery in the Cloud
In today’s economic environment, customers consider cloud platforms as a new variant of flexible and cost-effective solutions. Many organizations search for different ways to automate recovery of business data and the maintenance of working capacity of important applications during a power failure when the business process can be interrupted for a few hours (Cummins 2010). These days, the use of cloud platforms as solutions for disaster recovery meets the requirements of sustainability in organizations that cannot afford the maintenance of redundant sites or remote resources for important virtualized volumes. The purpose of the paper is to study the notion of the disaster recovery in the cloud as well as to evaluate its importance.
The Notion of Disaster Recovery
Cloud deployment may stop working because of the systematic shutdown of dependent services and basic infrastructure. In such circumstances, a business continuity plan initiates the process of disaster recovery. George Reese (2009) affirms: “Disaster recovery deals with catastrophic failures that are extremely unlikely to occur during the lifetime of the system” (110). This process usually involves employees of the operations department. It includes automated procedures to repeated activation of the application in the existing data center. It requires the user to move the data and the application service into a new data center. At the same time, media backups or replication are used (Bento 2012). Disaster recovery is a complicated task that usually includes possible loss of data and certain stoppage of operations. Susan Snedaker (2013) states that “Disaster recovery goes beyond just data backup and server protection and encompasses a complex mix of technologies, processes, and compliance issues” (465). For the enterprise, about two days are needed for the restoration of work. Moreover, the company suffers from such damages as the loss of profits and brand reputation (Grandinetti 2013). Organizations that turn to the providers of disaster-recovery-as-service for help assess the economy as a key advantage of using a public cloud for disaster recovery.
Advantages of Disaster Recovery in the Cloud
These days, data recovery in the cloud is extremely popular among various kinds of enterprises. However, before the advent of virtualization, restoration costs were very high as several DC, specialized software, and extensive network connections were needed. It was rather expensive to implement this in the material world (Sampathkumar 2015). Therefore, only the largest organizations could afford to spend money on recovery. Currently, virtualization has made the processes of backup and recovery easier by encapsulating the virtual servers in multiple files (Rao 2015). The data is easily transported, and cost of the procedure is reduced. Ganesh Deka emphasizes that “Some of the advantages of using cloud-based disaster recovery and business continuity are recovering workload from the cloud, where in the service provider will host the replicated system on the cloud, which will help manage disaster recovery in an easier manner” (37). Recovery procedures give data owners the variety of approaches to recovery.
To evaluate and comprehend the strategy of disaster recovery, a person needs to determine three terms – recovery point objective, recovery time objective, and recovery capacity objective. Recovery time objective represents the maximum time allowed for the recovery of application functionality. It is based on business requirements and associated with the importance of the application. Recovery point objective is a reasonable time window of data loss connected with the recovery process. For example, if recovery point objective constitutes about one hour, the data must be copied into the backup storage or replicated for at least every hour. After the resumption of the applications in another data center in backup data, data may be missing for one hour. Recovery capacity objective is defined by the functionality that should provide a backup system during recovery. The necessity of this indicator is caused by the fact that the power on the backup site may not be equal to the power on the main (Ackermann 2012). Thus, recovery occurs with some losses in functionality and performance.
From Backup to Full Synchronization
There are many options for implementing disaster recovery in the cloud. Which option to choose depends on various factors – the requirements for recovery speed, relevance recovered data, and others. The simplest version of a disaster recovery solution is the provision of copying data to a remote site. In this case, recovery time objective and recovery point objective receive a great significance. It should be understood that the backup is a periodic process rather than a continuous one. It means that if a backup is generated once every 24 hours, recovery time objective composes one day (Vacca 2013). If the main IT systems are destroyed, recovery time objective will also include not only the recovery of data from a backup copy but also the time of the recreation of IT systems. Then, recovery time objective could be several months, which is unacceptable. Therefore, with the help of only one backup, it is impossible to create a complete disaster recovery solution. Cloud backup is usually an integral part of disaster recovery (Vacca 2013). It provides backup of data to a remote site with minimal time on the organizational procedures. It is the first step in building a disaster recovery solution.
The second version of disaster recovery is large-scale. On the cloud backup site, computing power is created supported in standby mode. All the infrastructure settings are implemented there – installation of software, the configuration of security systems, and others. Business processes can move there at any time (Vacca 2013). It also implies the presence of a disaster recovery site copy of the data that does not lag from the primary one.
The most advanced version of the cloud-based disaster recovery implementation is the reservation of a full range of computing resources for the customer that are consumed by the IT systems. In this case, a copy of data located on the remote site is relevant at any time point and operates in an online mode (Vacca 2013). Creation and maintenance of efficiency of such a solution at the initial stage of the project requires a significant investment from the part of a provider.
Obstacles and Ways to Overcome Them
Sometimes, there are different obstacles in the implementation of disaster recovery in the cloud. Problematic connection channels between the cloud or the data center, in which the solution disaster recovery is deployed, and IT infrastructure of the customer can be the first obstacle. There are two ways to overcome it. The first one is to build a special fiber optic channel. The second way is to organize a dedicated connection channel from the ISP. Regardless of which of the methods the communication channel is provided, it is possible to bring information to the cloud safely. However, large public providers do not provide such technical features. They are mainly focused on standardizing services and allocate only a good channel of the public Internet. Nevertheless, nothing prevents to run data for disaster recovery over this channel, for example via VPN-tunnel. However, in this case, the rate of transmission will be too low, and because of the large distances between the systems, delays in transmission are inevitable. It is important to remember that technologies can supply dedicated communication channels to the public cloud in the interest of a particular customer’s work (Raj 2012). All restrictions are the artificial obstacles created by the provider.
Another obstacle is related to the standardization sought by cloud providers in the pursuit of economic efficiency. Frequently, the proposed options of standardization are in conflict with the interests of customers. One company may have the virtualization system from Microsoft, the other – from VMware. As a result, the provider may be in a difficult situation. The customer’s application systems must operate in such a virtualization tool that is absent in a standard set of solutions. However, as experience shows, it is realistic to implement the segment of the cloud on the technology to a particular company. In addition to standard RedHat KVM virtualizer, the cloud provider can offer solutions from Oracle and VMware work (Raj 2012). Another aspect that must be considered when creating a disaster recovery in the cloud is the inability of some IT systems to work on a flexible virtual platform. For example, sometimes, the customer places the entire IT infrastructure, which includes hard database, to one provider. This solution has a particular architecture. Moreover, it is so massive that can effectively operate only on the UNIX-based system. In the near future, the technology will not develop to the point that such solutions could be carried in a standardized cloud. Even if such radical changes occur, the customer will not agree to take the risk and change the existing scheme of work (Raj 2012). In this case, the solution is to connect hardware resources to the cloud dedicated to a particular customer.
Recovery of IT systems and services in the event of an accident represents a complex integrated solution that includes not only the necessary technologies but also the processes and procedures, professional training of employees, and their ability to act in an emergency situation. Not long time ago, disaster recovery was a highly expensive and complicated solution. It was often available only to the largest enterprises. Replicating the existing relations and infrastructure in the backup datacenter counted the greatest share of the price. This fact doubled the amount of investment in equipment. In addition, data replication mechanisms and software licenses also cost much. However, these days, due to virtual technologies, disaster recovery is possible not only as a complex individual decision but also as an affordable service for most enterprises.