Capacity Specification
Your deployment (the number of servers and the use case you have it for) determines your choice of NCache License, Edition, and Package. The use case helps determine what features you should use, thus, establishing what Edition you should use.
Important
For NCache Enterprise and Professional, you can move licenses from one environment to another by deactivating/reactivating them when needed.
After which, the most critical consideration is the number of vCPUs you will require.
Note
For assistance with the planning process, please contact Alachisoft Support.
As part of capacity planning (when you contact Alachisoft Support), you will have to provide information about your deployment. This includes the number of web servers you have, the number of cache servers you intend to use, and the hardware profile of your servers. You must ensure you have ample resources for deployment, and your primary focus should be on the following:
vCPUs: NCache is a highly multi-threaded software and takes full advantage of extra cores. The most common configuration for NCache is eight cores or higher. You may need more powerful processing if you have higher transaction loads or tonnes of data stored in NCache.
RAM: Ensure that the amount of memory you have available works based on your use case, as with more connected applications, there is more data, and more memory is required.
Network: You will require a network that minimizes latency and improves response time and cluster communication, as cache servers are incredibly chatty, and a fast network helps facilitate the constant communication between the application and the servers.
In most cases, a minimum of 8 vCPUs with 16 GB RAM is recommended, regardless of the size of your working set. However, you can use 4 vCPUs for 8 GB if they have a very simple use case.
Types of NCache Licensing Based on Capacity Specification
NCache licensing is vCPU based, where the license is bound to machine specifications (like the vCPUs, RAM, MAC addresses, IP addresses, machine names, etc.).
Note
NCache Servers refer to the boxes hosting the cache, and the term NCache clients refer to the boxes hosting your applications.
Based on these specifications, NCache offers two types of licensing, which have been discussed as follows:
Server-only (SO) Licensing
In server-only licensing, you only purchase NCache cache servers. These servers are licensed based on a combination of RAM & vCPUs. In this model, you get unlimited clients. This model is recommended if you have a large number of clients. While choosing this model, please ensure that your cache servers are mapped according to your plan.
Note
It is best if you follow the plans exactly as detailed. For example, if you opt for the 4GB plan, ensure that you have 4 GB available and 2 vCPUs.
| NCache Server-only (SO) Plans | Licensed RAM and vCPUs |
|---|---|
| 4GB SO-Plan | 4 GB, 2 vCPUs |
| 8GB SO-Plan (recommended) | 8 GB, 4 vCPUs |
| 16GB SO-Plan | 16 GB, 8 vCPUs |
| 32GB SO-Plan | 32 GB, 12 vCPUs |
| 64GB SO-Plan | 64 GB, 16 vCPUs |
| 128GB SO-Plan | 128+ GB, 32+ vCPUs |
Client-server (CS) Licensing
In client/server licensing, you must explicitly purchase and activate the clients and servers, compared to Server-only, where you only need to do this for the servers. This licensing model is recommended when you have a smaller number of clients.
| NCache Client-Server CS Plans | Licensed vCPUs (No RAM restriction) |
|---|---|
| 8 vCPU CS-Plan (recommended) | 1-8 vCPUs |
| 12 vCPU CS-Plan | 12 vCPUs |
| 16 vCPU CS-Plan | 16 vCPUs |
| 24 vCPU CS-Plan | 24 vCPUs |
| 48 vCPU CS-Plan | 32 vCPUs |
| 64 vCPU CS-Plan | 64 vCPUs |
Note
You can learn more about this and other licenses via NCache Pricing.
Note
NCache is very flexible in terms of deployment. You can deploy it on the cloud, on-prem, or using a containerized environment. Similarly, NCache lets you choose whether to opt for Windows or Linux.
Important
When deploying NCache, please ensure that everything (applications, cache servers, web servers, etc.) is on the same network, subnet, and IP range - avoiding latency.
