System Requirements¶
It is important to plan ahead and allocate resources based on the expected number and types of nodes in the simulations that the CML server will run. The CML application services have modest system requirements: when CML isn’t running any labs, the idle CML server uses only about 1 GB of memory and very little CPU. For system sizing, you should consider the maximum number and types of nodes that you expect to run simultaneously on your CML server. That is, for all labs that are running at the same time, count the number of nodes of each type in those labs, and add up the nominal amount of memory, CPU and at least minimum disk requirement used by each node. The CML FAQ on Reference Platforms page provides a table of the default resource requirements for each node type.
Note
CML is only fully supported when running on Intel processors. While the CML application services run equally well on systems with Intel or AMD processors, some of the bundled reference platform VM images only run on Intel processors. Support for running CML on non-Intel processors will be provided on a best effort basis.
CML Enterprise and Education offerings¶
Below are the recommended deployment options for CML Enterprise and Education offerings, each has its own advantages and disadvantages.
1. Installing CML on a Bare metal server¶
In this deployment type, CML is installed directly on a server, just like installing windows on a server, eliminating the need for a hypervisor.
CML is tested on Cisco M4, M5 and M6 models of the UCS C220 and UCS C240 rack servers with Intel processors. Full support for bare metal installations is only available on these server platforms.
CML is known to work on other vendor bare metal server installations with non-intel processors that supports nested virtualization, however support on these servers will be on best effort basis.
Advantages:
Eliminates the need to buy a hypervisor license, as CML is directly installed onto to the physical server.
Installing CML on bare-metal provides marginal improvements in performance as it removes a layer of virtualization.
Disadvantages:
Limited backup options.
No flexibility to add more appliances to the host server if CML does not require all of the host resources.
2. Installing CML on ESXI¶
In this deployment type, CML is deployed as a VM in ESXI. Processor must support VTx and EPT virtualization technologies.
Requires ESXI Version 6.5 or later
While the CML application services run equally well on systems with Intel or AMD processors, some of the bundled reference platform VM images only run on Intel processors. Support for running CML on non-Intel processors will be provided on a best effort basis.
Advantages:
Installing CML on ESXI facilitate easier management of CML VM. For eg: adjusting resource allocations, adding network, adding storage etc.
Provides ability to take backup of CML VM via snapshots.
Non-Intel processors can be used provided it supports VT-X and EPT technology.
Disadvantages:
Incurs additional cost of purchasing the hypervisor.
Performance may get affected depending on footprint of other VMs in the same host.
3. Installing CML on cloud platforms (AWS and Azure)¶
For a fully functional deployment of CML on cloud platforms, selecting instances with support for nested virtualization is crucial, as many CML node types may not operate correctly without this feature. While some compute-optimized instances may provide limited usability, this approach is experimental and typically not supported, potentially leading to reduced functionality.
Advantages:
CML can be quickly set up on cloud platforms, enabling immediate start of network simulations.
Cloud deployment of CML removes the necessity for personal hardware investment and maintenance.
Cloud resources can be easily scaled to accommodate the demands of complex network simulations.
CML can be accessed from anywhere, making it convenient for teams that are distributed geographically.
Cloud platforms provide various services that can be integrated with CML for enhanced functionality, such as automated backups and advanced networking options.
Disadvantages:
Running CML on cloud platforms can lead to higher operational expenses.
Some cloud platforms may lack out-of-the-box automation features for CML, requiring manual setup and management.
Depending on the cloud provider and the complexity of the deployment, there may be limited support options for CML-specific issues.
Managing CML in the cloud requires familiarity with both the CML application and the cloud platform’s services.
CML Personal and Personal plus offerings¶
Below are the recommended deployment options for CML Personal and Personal plus offerings.
1. Installing CML on Windows based System¶
CML is installed on a Windows machine with VMware Workstation/Player.
Officially supported only on systems with Intel processors. Known to work in AMD systems.
2. Installing CML on a MAC System¶
CML is installed on a MAC machine with VMware Fusion.
Officially supported only on systems with Intel processors. CML does not work in M series chips.
Installing CML in Cloud¶
Both CML Enterprise and Personal offerings support cloud deployments in AWS and Azure. Please refer CML github page for detailed instructions for cloud installations.
CML Server / VM Minimum Requirements¶
Important
The values listed below are the minimum recommended requirements for CML 2.x. Using these minimum values will restrict the number of nodes that you can simulate and could impact overall system performance. If you are deploying CML as a virtual machine, the memory, CPU, and disk size values are the resources that must be allocated to the CML VM, not the values for the host OS.
System Resource |
Minimum Requirements (default configuration) |
|---|---|
Memory |
8 GB |
CPU |
4+ (physical) cores, Intel processor Processor must support VTx and EPT. |
Network |
1 Interface |
Hard Disk |
32 GB or more |
Non-personal CML product offerings, such as CML-Enterprise and CML-Education, are tested on both M4 and M5 models of the UCS C220 and UCS C240 rack servers. Full support for bare metal installations is only available on these server platforms.
Supported Software¶
The CML web-based user interface requires an HTML5-capable web browser. The UI is tested with both Chrome and Firefox and is supported on those web browsers.
The following hypervisor versions are supported for the CML VM-based deployment option. The OVA file’s virtual hardware version is 10.
Virtualization Platform |
Version |
|---|---|
VMware Workstation |
15 or later |
VMware Fusion Pro |
11 or later (Intel models only) |
VMware Player |
15 or later |
VMware ESXi |
6.5 or later |