October 18, 2017
How to Increase VM Backup Speed to Its Maximum
Backup speed becomes a serious issue when your business operates 24/7 with hundreds of virtual machines up and running. You can imagine that it gets all the more serious when the number of VMs increases to thousands. From this perspective, boosting your VM backup speed becomes a vital element in keeping up your virtual infrastructure resilience.
Of course, not everything depends on the agility of backup software. Your virtual environment may also suffer from insufficient network bandwidth creating bottlenecks between the source data storage and the backup software or the backup software and the destination data storage, or low write speed of the target storage.
Anyway, while the above relates mostly to resources availability vs. cost, modern backup solutions can provide you with the maximum possible VM backup speed in the given virtual environment. In this regard, legacy VM backup solutions leave much to be desired, lagging far behind the ever-growing demands of the virtualized environments. Let’s review features you should be interested in.
Full Synthetic Approach to Storing Backups
In contrast to the traditional backup types, such as full, incremental, or differential, the synthetic backup emerged quite recently and does not require performing periodic full backups. A full backup is created only once, and the following backup jobs are forever-incremental. Solutions using the full synthetic approach to creating and storing backups use CBT and/or RCT technologies allowing to quickly identify changed data blocks. The full synthetic data storage mode allows saving changes to the backup repository as increments, and recovery points are created with references to these changes. Other techniques like data deduplication and compression reduce the data size and update recovery point references to the deduplicated data blocks.
The example below illustrates this process and shows how changes (or increments) occurred within three days in the source VM are then stored in the backup repository.
The changed data blocks and three recovery point references in the backup repository allow to restore the full VM to its state as of any of these three days.
Thus, using the full synthetic mode of storing backups allows to drastically reduce both the size of the backup repository and network workload, and, as a result, speed up the backup process.
Elimination of Swap Data from VM Backups
Swap files and partitions are essential for keeping up operation of applications in case they need more RAM than is physically available. However, having the backup system loaded with processing, transferring, and storing temporary swap files and partitions, which are unnecessary for recovery purposes, takes a toll with regards to VM backup performance.
The Skip Swap Data feature available in modern VM backup solutions allows for automatic elimination of swap files and partitions from VM backups, thus reducing the backup size and, most importantly, increasing the speed of the backup process.
LAN-free Data Transfer
When VM backup is performed, large amounts of data are transferred through the network, which can affect the speed of both the backup process and business-critical operations. However, it is possible to optimize backup traffic, and one of the best ways to do this is to use LAN-free data transfer. This technology is based on the Direct SAN Access and HotAdd techniques.
Direct SAN Access mode lets the backup solution read and transfer data over Fiber Channel or ISCSI directly from a SAN storage device, which increases the speed of backup and replication jobs and greatly improves VM backup performance.
Alternatively, the HotAdd transport mode enables reading VM data directly from VM datastores through the storage I/O stack bypassing the host’s TCP/IP stack. A snapshot of a source VM is “hot added” as a virtual disk to the VM that runs a backup application. This feature also helps improve backup speed.
To make VM backup jobs run faster, you can also reduce the network load and minimize backup windows. With the Network Acceleration feature onboard, you can do all of the above, while boosting the speed of your backup jobs by up to 2X in LAN and WAN networks. This technique is especially efficient for remote VM backup repositories – offsite or in the cloud.
VM Backup Job Scheduling
Running multiple backup jobs concurrently in large environments can critically slow down your infrastructure performance due to temporary, but intensive workload increase, stressing your network resources. To mitigate these peaks, you need to carefully schedule your backup jobs.
This implies that two aspects should be taken care of:
- Scheduling backup jobs to run at times when the workload on the resources is minimal (it may be a rather difficult task with large environments running transactional applications, when the activity hardly drops any time of the day);
- Careful selection and grouping of backup jobs to keep the time slots for backup jobs as narrow as possible.
The best VM backup solutions incorporate convenient and practical tools like Calendar Dashboard for easy and flexible scheduling of your backup jobs.
Besides resolving numerous issues related to VM backup, replication, and recovery, modern VM backup solutions have to provide seamless integration with the rest of the infrastructure, as well as ensure smooth and well-balanced performance of the whole data protection system. Here come the industry leaders to meet the challenge. Among them, NAKIVO Backup & Replication, which is the only solution on the market to offer a full package of VM backup technologies boxed in a single product. With its outstanding suite of features, NAKIVO Backup & Replication provides VM backup, replication, backup to cloud, global data deduplication, instant VM and object recovery, backup copy, and screenshot verification. The product offers you all features necessary to increase VM backup speed, including those examined in this blog post.