The evidence continues to mount, that vRanger Pro 4.5 is faster for backup and recovery than other image-based protection offerings. The data shared here-in comes from a new Vizioncore customer evaluating vRanger Pro 4.5 against a competitive product.
The key finding: Make sure that you test with real-world numbers of VMs, and not just one at a time, when evaluating backup products. With 33 VMs rather than just one, our new customer found that:
Purpose of test:
Validate which product has better performance. Also, compare the performance of LAN-based versus LAN-free backup.
Environment and Backup set:
ESX HA & DRS cluster with 3 x Pe2900 hosts, 1 x Md3000i ISCSI SAN with 2 LUNs, 1 x Pe2900 backup host. All servers connected with ISCSI via 2 Dell ISCSI switches, 2 subnets. All connections gigabit full duplex. ISCSI jumbo frames NOT turned on, no LAN jumbo frames either. Destination for backup data 1 x MD3000i ISCSI SAN.
Full Backup of 33 VMs (all running VMs - 31 Win, 2 Linux)
2.7 TB VMDK thin provisioned
LAN-Based backup result:
Vizioncore vRanger 4.5
Competitor config 1
Competitor config 2
Backup window / performance
557 GB, 116 files 64 folders
554 GB 1 file
575 GB 1 file
Don’t use fiber/ISCSI (direct-to-target)
Enable VMware quiescing
Don’t check free space
VMware vStorage API Network mode, no traffic encryption
No auto backup integrity checks
Enable VMware Tools quiescence
Inline deduplication optimal
Use CBT, enable for all
Disable VMware Tools quiescence
Result: vRanger Pro 4.5 is 3-4x times faster than competitor in LAN-based backup
LAN-free backup result:
631 GB 116 files 64 folders
544 GB 1 file
Common backup options
LAN free full backup (“use ISCSI, network if fails” and “VMware Storage API SAN mode with network failover”)VMware Tools Quiescence
Compress each VM archive
Inline deduplication, level “best”
Result: vRanger Pro 4.5 is 1.75x times faster than competitor in LAN-free based backup
Why LAN-based backups faster with vRanger
Depending of your environment the speed can differ. But, in most cases, LAN-based backups are faster than LAN-free, when you run big backup jobs across multiple hosts.
This is because you get higher bandwidth over the network to the repository compared to fiber network.
Example: 10 ESX hosts with 1 Gbit/s network gives you 10 Gbit/s backup pipe compared to 4 Gbit/s if you have Fiber Channel.
We have seen customers reporting 1GB/s in backup throughput when using 10Gbit/s Ethernet and 10Gbit/s repository. That´s not something you will see from others who claim super-lightning-fast backups!
Comment from tester:
“With small backup sets or single VMs, [we found that the] competitor is often faster and has better compression. In my first test rounds where I tested backup of less than 10 VMs they beat [vRanger Pro]. It was when I started real backup tests with 2,7 TB VMDK size and 33 VMs that your multithreading wins the day. That’s why I initially was very tempted to buy competitive solution because of better speed in that test and some GUI things I like in their solution. So I’m glad I did a real production tests too J .”
What makes the difference?
vRanger 4.5 has an intelligent and scalable backup engine called Data Protection Platform (DPP) that allows multiple backups to run at the same time. This also applies to recovery jobs, which means that multiple backup AND recovery jobs can run at the same time.
The Direct-to-Target capabilities built into DPP enable data to be sent directly to the backup target repository.
Active Block Mapping (ABM) is Vizioncore patent-pending technology that skips processing of deleted files/blocks and avoids blocks with zeros.
The testing organization purchased vRanger Pro and became a new Vizioncore customer.