How Can I Improve My Disk Performance?

Method

Application Scenarios

Changing the EVS disk type

If your disk (for example, an Ultra-high I/O disk) can no longer provide the IOPS or throughput required by service growth, switch to a higher-performance disk type. Doing so can significantly increase the I/O upper limit and reduce response latency. This method is suitable for applications that require high storage performance and are facing rapid growth in service scale or access traffic.

If you are using General Purpose SSD V2 or Extreme SSD V2 disks, you can adjust their performance configuration to further improve performance.

Expanding the disk capacity

If your face both performance bottlenecks (IOPS and throughput) and insufficient capacity, you are advised to expand the disk capacity. For some disk types, capacity and performance scale linearly, so expanding capacity can also increase baseline IOPS and throughput. This method is especially suitable for mission-critical services with continuously growing data and strict requirements for capacity and concurrency.

Using LVM to manage EVS disks

Logical Volume Manager (LVM) is an ideal choice when you need flexible storage resource allocation and want to overcome the performance limits of a single disk. By stripping data across multiple physical volumes, LVM enables parallel read and write operations, significantly improving the I/O throughput of EVS disks. This method is especially suitable for mission-critical services that require high concurrency and low latency, such as databases and multi-thread applications.

Using EVS disks to create RAID arrays

If you need to improve IOPS and throughput while ensuring data redundancy, you can create RAID arrays. RAID 0 is suitable for scenarios that require high performance but do not require data redundancy; however, if a single disk fails, the entire array becomes unavailable. RAID 1 is suitable for scenarios that require high data redundancy but do not require high performance. RAID 10 combines the advantages of RAID 0 and RAID 1, making it suitable for scenarios that require both high performance and high data redundancy.