**6.8 Disaster recovery-as-a-service (DRaaS) challenges**

In this section, we investigate some common DRaaS challenges in cloud environments.

#### *6.8.1 Cost*

One of the main factors to choose cloud as a DRaaS is its price. CSPs always seek cheaper ways to provide recovery mechanisms by minimizing different costs.

#### *6.8.2 Replication latency*

DRaaS depends on the replication process to create backups. Current replication strategies are divided into two categories: synchronous and asynchronous. However, they both have certain advantages and disadvantages. Synchronized replication, guarantees excellent RPO and RTO, but is dear and might affect system performance because of over-optimization. This problem is incredibly bad for multi-tier web applications because it can greatly increase the trip Time (RRR) between the first and backup sites. On the choice hand, the backup model adopted with asynchronous replication is cheaper and also the system has fewer problems, but the standard of the DRaaS is reduced. Thus, the transaction between costs, the performance of the system and replication latency is an undeniable challenge in cloud disaster solutions.

#### *6.8.3 Data storage*

Business database storage is one of the problems of enterprises that can be solved by cloud services. By increasing cloud usage in business and market, enterprises need to store huge amounts of data on cloud-based storage. Instead of conventional data storage devices, cloud storage service can save money and is more flexible. To satisfy applications and to guarantee the security of data, computing has to be distributed but storage has to be centralized. Therefore, storage a single point of failure and data loss are critical challenges to store data in cloud service providers [17].

#### *6.8.4 Lack of redundancy*

When a disaster happens, the primary site becomes unavailable and the backup site has to be activated to replace the primary site. It is a serious threat to the system. This issue is temporary and will be removed once the primary site is recovered.

#### *6.8.5 Failure*

The early detection failure time significantly affects the recovery time of the system, so it is important to detect and report rapid and accurate DRaaS failure. On the other hand, in many backup sites there is a big question: How to separate network failures and service interruptions.

### *6.8.6 Security*

As mentioned earlier, disaster can be created by nature or can be man-made. The cyber-terrorist attack is one of the most man-made disasters that can occur for a variety of reasons. In this case, protecting and restoring important data will be the focus of the DRaaS programs other than system restoration [18].

#### **6.9 Disaster recovery-as-a-service (DRaaS) solutions**

In this section we discuss some DRaaS solutions that can address the problems and challenges raised by cloud-based DR.

#### *6.9.1 Local backup*

A solution to the dependency problem has been proposed in [19]. A Local Backup can be deployed on the side of customers to control data and to get backup of both data and even complete application on local storage. Local storage are often updated through a secured channel. By this technique, migration between CSPs and migration from public to private clouds, and from private to public clouds is possible. In the event of a disaster, local backups can provide the services that used to be provided by the CSP.

#### *6.9.2 Geographical redundancy and backup (GRB)*

Geographical Redundancy can be used in a traditional context. Two cloud zones mirror each other [17]. If one zone gets down, then the zone will be on and provide the services. A module monitors the zones to detect disaster.

Another research [20] has been proposed proposes a method to select optimal locations for multiple backups. The number of places is decided based on the nature of the application and service priority. Distance and bandwidth are two factors to settle on the simplest sites. However, this work neglects some critical factors such as the capacity of mirror sites and the number of node sources that can be hosted in each location.

#### *6.9.3 Inter-private cloud storage (IPCS)*

According to the Storage Networking Industry Association (SNIA), at least three backup locations are necessary for business data storage. Users' data should be stored in three different geographical locations: Servers, Local Backup Servers (LBS) and Remote Backup Server (RBS) [21].

#### *6.9.4 Resource management*

Hybrid clouds consist of many different hardware and software. In cloud-based enterprises, all business data are stored in storage resources of the cloud. Therefore, data protection, safety and recovery are critical in these environments. Data

protection is challenged when the data that has been processed at the primary host has not been stored in the backup host yet. There are three solutions for data recovery purposes [22]:

