Overview

Dedicated Distributed Storage Service (DSS) provides dedicated physical storage resources. With data redundancy and cache acceleration technologies, DSS delivers highly reliable, durable, low-latency, and stable storage resources. CCE allows you to mount DSS storage volumes to containers.

DSS Performance

The key indicators to measure the performance of DSS storage pools include IOPS, throughput, and I/O latency.

  • IOPS: the number of input/output operations per second

  • Throughput: the amount of data read from and written into a storage pool per second

  • I/O latency: the minimum interval between two consecutive I/O operations

Table 1 DSS performance specifications

Parameter

High I/O

Ultra-high I/O

IOPS

1500 IOPS/TB

8000 IOPS/TB

I/O latency (single queue, 4 KiB data blocks)

1 ms to 3 ms

1 ms

Typical application scenarios

Common development and testing

  • Transcoding services

  • I/O-intensive services

    • NoSQL

    • SQL Server

    • PostgreSQL

  • Latency-sensitive services

    • Redis

    • Memcache

Application Scenarios

DSS supports the following mounting modes based on application scenarios:

  • Using DSS Through a Static PV: static creation mode, where you use an existing volume to create a PV and then mount storage to the workload through a PVC. This mode applies to scenarios where disks are available.

  • Using DSS Through a Dynamic PV: dynamic creation mode, in which you do not need to create disks beforehand. Instead, specify a StorageClass when creating a PVC. Then, a disk and PV will be created automatically. This mode applies to scenarios where no disk is available.

  • Dynamically Mounting a DSS Disk to a StatefulSet: available only for StatefulSets. In this mode, each pod is associated with a unique PVC and PV. After a pod is rescheduled, the original data can still be mounted to it based on the PVC name. This mode applies to StatefulSets with multiple pods.