Retired local SSD disks

Note the following when you use local SSDs:

• The storage is from a local SSD on a single server. This creates a single point of failure threat. Implement data redundancy at the application layer to ensure data availability.

• You cannot upgrade or downgrade the CPU, memory, or local SSDs after purchase.

Local SSDs use the local SSDs on the physical server on which the instance resides and have the following features:

• Low latency: Provides microsecond-level access latency.

• High random I/O: Provides up to 12,000 random IOPS.

• High throughput: Provides up to 300 MB/s of I/O throughput.

• Limited data reliability: Local SSDs are on a single physical server. Data reliability depends on the reliability of that server, which creates a single point of failure threat. Implement data redundancy at the application layer to ensure data availability.

• Up to 800 GB capacity: A single local SSD provides up to 800 GB of storage.

• Cannot be attached or detached: Because the disks are local to the physical server, you cannot attach or detach them independently.

• You cannot upgrade or downgrade the CPU, memory, or local SSDs after purchase.

Local SSDs provide low latency, high random I/O, and high throughput. They are suitable for I/O-intensive distributed applications that have built-in redundancy and do not require high data reliability. Suitable scenarios include the following:

• Distributed applications: I/O-intensive applications such as NoSQL, Massively Parallel Processing (MPP) data warehouses, and distributed file systems. These applications have built-in data redundancy. Local SSDs provide the low latency, high random I/O, and high throughput that these applications need.

• Large-scale online application logs: Large online applications generate large amounts of log data. These applications require high-performance storage but do not have high requirements for data reliability.

• Swap partitions for instances: You can use swap space in Linux when an application requires more memory than is allocated. When swap space is enabled, the Linux system frequently moves memory pages from physical memory to the swap space. The swap space can be a dedicated partition or a swap file. This process frees up space in physical memory for pages that require fast access.