SoftNAS Instance Size Guidance
SoftNAS® offers the broadest range of instance sizes and region availability on both Amazon Web Services™ and Microsoft® Azure™. It’s important to select the right instance size to configure a storage solution that is the right combination of performance and price for your use case. General guidance is provided below to help you select an instance size for your workload.
How to Select an Instance Size
- Select a performance category that most closely describes your expected workload.
- Review the resource properties for the chosen category to confirm they match your expected level of instance resources.
- Use the tables below to help you select an instance from the chosen performance category.
If unsure of your final selection, start with a larger instance size initially, deploy or simulate your production workload, then monitor and observe your results for 30 to 45 days
- If your actual workload peaks are less than 50% CPU and network usage, then you may be able to reduce the instance size such that the peaks do not exceed 70%.
- If CPU or network usage reaches 75% or higher, then an increase in instance size may be appropriate.
Due to unique environments and use cases, we recommend continued monitoring of your SoftNAS instances using the SoftNAS StorageCenter™ and your cloud provider’s virtual machine metrics.
To help you select an instance for your SoftNAS deployment, we have created 4 performance categories (Developer, General Purpose, High Performance, Extreme Performance). The categories offer different levels of price and performance based on 3 key properties of the instance:
- Network Speed – The speed of your network can be a primary consideration when using network attached storage.
- RAM – SoftNAS uses 50% of available RAM for level 1 Read cache.
- Disk Read Cache/Write Log – EBS or SSD disks can be used for level 2 Read cache and/or Write log. This property must be configured after the instance is launched.
While there are other factors to consider, the above 3 instance properties are excellent starting considerations when selecting an Azure VM or AWS instance for SoftNAS.
For more information on SoftNAS Performance Consideration, click here.
Developer level workloads are suited for DevOps, Administrators, Architects and other team members looking to learn more about SoftNAS without incurring the higher costs of full production-level instances. Developer level includes Free Tier micro instances where available. Developer category instances are not suitable for performance targeted Proof of Concepts or production environments. High-availability (HA) is not supported for this category of instances.
- Network: 100Mb (12MB/sec) or less
- RAM: 4GB or less
- CPU (Cores or vCPUs): 1 or less
- Disk Read Cache/Write Log: AWS EBS or up to 160GB for select instances
Entry level instances are good choices for workloads with basic requirements for network bandwidth, caching, deduplication, encryption and RAID. This instance category is a good starting point for Proof of Concept trials and general workloads with a lower price over performance.
- Network: 100Mb/sec (12MB/sec) or 1Gb/sec (120MB/sec)
- RAM: 8GB to 28GB
- CPU (Cores or vCPUs): 2 to 4
- Disk Read Cache/Write Log: AWS EBS or up to 200GB SSD for select instances
Demanding workloads that are read intensive will benefit from increased memory-based read cache included with High Performance instances. Additional CPU resources will provide better performance for workloads that require more extensive use of deduplication, encryption, compression and/or RAID. Read-intensive workloads should leverage the use of Level 2 Read Cache using direct-attached SSD or EBS Provisioned IOPs storage (EBS PIOPs for AWS only).
- Network: 1Gb/sec (120MB/sec)
- RAM: 28GB to 64GB
- CPU (Cores or vCPUs): 8 to 16
- Disk Read Cache/Write Log: AWS EBS or up to 400GB SSD for select instances
Extremely demanding workloads requiring thousands of concurrent connections, maximum IOPs, and throughput should use Extreme Performance instances. Level 2 Read Cache and Write Logs using direct-attached SSD or EBS Provisioned IOPs storage (EBS PIOPs for AWS only) should be leveraged for extreme workloads.
- Network: 10Gb/sec (1.2GB/sec) or 20Gb/sec (2.5GB/sec)
- RAM: 56GB to 256GB
- CPU (Cores or vCPUs): 16 to 64
- Disk Read Cache/Write Log: AWS EBS or up to 1000GB SSD for select instances
AWS Instance Sizes
|Performance Categories||Suggested Instance¹||Other Considerations²|
Microsoft Azure Instance Sizes
|Performance Categories||Suggested Instance||Other Considerations|
¹ The Suggested Instance is the default choice recommended for each Performance Category.
² Other listed instances may be appropriate choices based on the amount of RAM, CPU resources, Local vs. Network Attached disks and pricing considerations.