There are three primary data replication processes;
- Array-Based Replication
- Network-Based Replication
- Host-Based Replication
Here, we will discuss the simplest and most agile process of the three i.e. the Host-Based Replication.
The Host-Based Replication is a process of duplicating data from one site to the target place and works with any storage type such as NAS, DAS, and SAN.
A file system filter driver (the replication agent) is installed onto the server to be replicated.
The agent operates on the I/O storage system’s traffic and replicates it to its destination. Its software can be installed within an OS or a virtual machine, locally.
It is critical for an organization to keep an uninterrupted data flow and security throughout. Host-based replication software usually includes compression, encryption, throttling, and failover options too.
A host-based replication compared to the other to the data replication process is relatively inexpensive and has a simpler process.
It is perfectly reasonable to adapt host-based replication software if an organization is seeking no impact on the operation after a disaster.
Host-based replication is a storage-agnostic and uses server resources. The other two data replication methods- array and network are OS-agnostic, a characteristic that a Host-based Replication methodology lacks.
Furthermore, the host-based implementation adds fewer points of overhead to the servers’ CPU application. That is why the businesses with higher transaction volumes requirement tend not to prefer the host-based replication.
A host-based replication is an excellent tool, especially for SMBs. That is because the process is easier to monitor and provide support, while the server numbers are low. Once the number of nodes increases the scalability of the process becomes a challenge.
The initial and target storage devices need to be active to carry out a host-based replication process. Furthermore, unlike the array and network-based replication, it is more prone to viruses, resource shortages, and application failure.
The ease of installation, inexpensiveness, simplicity, and customization has made the host-based replication the most widely deployed replication method for the SMBs.
The same could not be said for the network-based replication, which has a significantly higher cost and complexity in processing than the former.
Host-based replication provides small and medium businesses, the enterprise-level data protection and disaster recovery while keeping the site cost to a minimum.
As the host-based replication agent needs to be installed and managed on each of the servers, therefore it is less scalable for the networks with a larger no. of servers.
The existing backup infrastructure needs to be taken into consideration when evaluating host-based replication products.
We’re seeing a trend of backup application vendors adding replication-based data protection to their suites.
They view replication as a feature within their backup suites rather than a standalone product. The majority of backup software vendors already offer host-based replication options for their products:
SIOS Technology (SteelEye DataKeeper) is an affordable host-based replication solution provider. It is a lightweight software with real-time data replication and protection applications in physical, virtual, and cloud environments.
It has a simple installation process with quick replication across LAN and WAN with minimal bandwidth. It keeps the performance impact of data replication on your application servers and network under check.
Veritas Volume Replicator
VVR, a host-based replication product coming from Veritas is basically an additional option that falls under VxVM (a storage management system).
Veritas Volume Replicators support both types of replication modes i.e. synchronous as well as asynchronous replication.
Let’s dive into the components of VVR to better understand it:
– Replicated Volume Group: RVG is essentially a group of volumes i.e. the data meant for replication.
– Storage Replicator Log: This is a chain that consists of all the operations/writes that are meant to be sent from primary to secondary.
– Replication Link: Various attributes that help correspond writes between primary and secondary RVGs.
– Data Change Map: Meant to keep a track of writes by creating a bitmap as and when SRL is overburdened.
– Replicated Data Set: It is a concept that helps put different RVGs in the primary in context with their counterparts in the secondary.
vSphere Replication is another one of the host-based replication systems. This replication system is in sync with VMware vCenter Server and VMware vSphere Web Client.
Unlike VVR, vSphere operates under one mode of replication only i.e. asynchronous replication of virtual machines.
One of the beneficial things that come with vSphere Replication is that the means of storage is not a concern. It can be virtual or your traditional means of storage.
Below we have listed out some of the key features and benefits:
- Monitoring this system isn’t a concern due to it being in sync with vSphere Web Client.
- Protection and recovery of data is managed with the same as well as across other sites.
- Doesn’t depend on the operating system of the virtual machine.
vSphere Web Client basically sets up the process of replication for a VM. Once this is done, vSphere Replication puts the source and target machines in sync.
The time in putting the two units in sync completely depends on the amount of data that is up for replication. A duplicate of the virtual machine disk is created that is meant to be transferred.
After the initial syncing is set up, a constant track of replicated data is kept (referred to as lightweight delta syncs).
Host-based data replication has now become a pivotal part of data protection and disaster recovery for most businesses.
Even with so many more advanced options for data replication, host-based methodology continues to be a widely used data solution because of its easy implementation, integration, and cost-effectiveness.