This afternoon, I was pleased to be on a webcast with Redmond Magazine titled “Rethinking Data Protection,” where we discussed issues related to hybrid data protection and high availability. Arcserve was a sponsor for this webcast along with Company ”Z.” If you missed the live webcast, you can view the recorded webcast at this link.
As discussed in the webcast, choosing the right backup and high availability solution for your organization can be a tricky, if not daunting task. In this blog, I would like to dive deeper into the key difference of Arcserve’s High Availability as compared to vendors that specialize in virtualization, like Company “Z”, and also open the conversation to a the broader picture of data protection.
Arcserve High Availability (HA) is an option of Arcserve UDP. It is an agent-based, real-time replication and high availability solution that allows for replication/HA from virtual and physical environments, or a mixture of both (P2V, V2P, V2V and V2P). It can be configured to automatically failover or failover through user intervention, which enables RPO/RTOs of mere seconds. It complements other availability features available in UDP.
Alternatively, Company “Z” is a hypervisor-based, agentless VM replication solution that requires a user to log into the site Company “Z” Virtual Manager and press a failover button (there isn’t a high availability or automatic failover solution).
Key elements worth consideration are:
- Arcserve HA requires a minimum of only one control service engine with RHA manager to two servers, if redundancy is required for scenario management. In contrast, Company “Z” requires a Company “Z” Virtual Manager on each site, a Company “Z” Virtual Appliance on each virtual host and a Company “Z” Cloud Manager in order to manage multi-tenancy. This can range from a minimum of four to five servers and upwards, depending on the amount of hosts.
- With Arcserve HA, setup includes a myriad of failover options. It enables manual file and application failovers, full system high availability or application high availability where failover can be manual or automated with integrated DNS changes to the local DNS server. On the other hand, Company “Z” provides simplistic Virtual Protection Groups to be created but doesn’t offer configuration flexibility with the exception of pointing to a secondary site Company “Z” Virtual Manager and some Re-IP configuration for the disaster recovery site.
- In a replication scenario with Arcserve HA, a live disaster recovery VM or server is used to receive replicated files. With application HA, the server does require the application to be set up and configured, which entails more running computer resources on the disaster recovery site – however this has its added benefits. Alternatively, Company “Z” replicates VM files at a block level to the disaster recovery site pool, and only creates and powers a VM onto the disaster recovery hypervisor once failover is initiated (regardless of whether it’s a test or real scenario). This being the case, Company “Z” cannot guarantee RTOs.
- Arcserve HA includes data rewind (similar to Journal History) that enables the user to rewind changes made to an application or the OS files – whereas Company “Z” only offers this capability at a hypervisor block level.Arcserve HA offers ‘Assured Recovery’ which enables automated HA and disaster recovery testing, and unlike Company “Z”, tests the application in the guest OS as well as data consistency with the master server. Alternatively, Company “Z” does not test application or data consistency within the guest OS of the protected VM. Testing is everything in BC/DR.
I hope that you now have a deeper understanding about our comprehensive approach to HA.
It’s more than HA
Our HA module is an option of our Arcserve Unified Data Protection solution that has taken the market by storm in the past year and a half. One final consideration: not every workload necessitates true high availability. Many systems and data sets can withstand RPOs and RTOs of a few minutes. Our UDP solution is very flexible in allowing you to create data protection plans that can combines multiple techniques such as deduplication, replication, snapshots, virtual standby, instant VM, tape archive etc. These plans are literally like an RPO/RTO dial – easy to create and customize, easy to deploy and match to your business needs. When deploying data protection, is the right approach to simplify under one roof with coherent processes that give you full RPO/RTO control, or is going for a narrow/specialized approach more efficient? 20,000 thousands customers in the past 1.5 years would confirm that the unified approach of Arcserve UDP not only provides more comprehensive data protection, but a better high availability solution.