High Availability Deployment Requirement

To deploy NovoOne adopting high availability deployment architecture, you need to know the requirements of server specification, operating system and network.

Note: Deployment requirements described in this topic apply to NovoOne deployed on on-premises virtual machines (VMs).

Server specification

With High Availability Deployment Architecture, NovoOne delivers scalable concurrent call limits via node scaling, with the following matching server specifications.

Note:

With the server specifications in this section, the ratio of the concurrent call limit to the user registration limit is 1:50.
For example, NovoOne deployed with the server specifications for 100 concurrent calls supports up to 5000 user registrations.
For deployments supporting more than 1000 concurrent calls, contact Yeastar Support for further recommendations.
When deploying two sets of NovoOne clusters across different regions for disaster recovery, double the server specifications specified in this section.

Specifications reference for 100 concurrent calls
Specifications reference for 200 concurrent calls
Specifications reference for 300 concurrent calls
Specifications reference for 500 concurrent calls
Specifications reference for 800 concurrent calls
Specifications reference for 1000 concurrent calls

Minimum server specifications for 100 concurrent calls

This deployment consists of 8 nodes - 6 worker nodes for Kubernetes Cluster and 2 load balancer nodes for Load Balancer Cluster. Node workloads are as follows.


Cluster	Node	Description
Kubernetes	Worker-1	Host core service components (e.g., Edge, HTTP Proxy, and Database), allowing flexible pod scheduling for efficient resource utilization.
	Worker-2
	Worker-3
	Worker-4
	Worker-5	Dedicated to block storage component to provide persistent storage for prompts, call recordings, and other data.
	Worker-6
Load Balancer	LB-1	Operate in an active-standby mode to ensure high availability through failover. The active LB node performs health checks and distributes traffic.
Load Balancer	LB-2

The minimum server specifications for each node are as follows:


Node	CPU	CPU frequency	vCPUs	Memory	Storage				Number of fixed public IP address
Node	CPU	CPU frequency	vCPUs	Memory	System disk	Data disk	Disk IOPS	Disk throughput	Number of fixed public IP address
Worker-1	x86‑64	2.5 GHz	8	32 GB	150 GB	/	3000	125 MB/s (250MB/s recommended)	1
Worker-2			8	32 GB	150 GB	/			1
Worker-3			8	32 GB	150 GB	/			1
Worker-4			8	32 GB	150 GB	/			1
Worker-5			4	8 GB	50 GB	> 500 GB			/
Worker-6			4	8 GB	50 GB	> 500 GB			/
LB-1		/	2	4 GB	/	/	/	/	1
LB-2		/	2	4 GB	/	/	/	/	1
Total	/	/	44	152 GB	700 GB	＞ 1 TB	/	/	5

Important: The Data disk on each of Worker-5 and Worker-6 must have sufficient additional disk capacity to store prompts, call recordings, and other data.

Minimum server specifications for 200 concurrent calls

This deployment consists of 11 nodes - 9 worker nodes for Kubernetes Cluster and 2 load balancer nodes for Load Balancer Cluster. Node workloads are as follows.


Cluster	Node	Description
Kubernetes	Worker-1	Host core service components (e.g., Edge, HTTP Proxy, and Database), allowing flexible pod scheduling for efficient resource utilization.
	Worker-2
	Worker-3
	Worker-4
	Worker-5	Dedicated to media processing component to increase processing capacity.
	Worker-6
	Worker-7	Dedicated to monitoring components to isolate monitoring workloads from core call services.
	Worker-8	Dedicated to block storage component to provide persistent storage for prompts, call recordings, and other data.
	Worker-9
Load Balancer	LB-1	Operate in an active-standby mode to ensure high availability through failover. The active LB node performs health checks and distributes traffic.
Load Balancer	LB-2

The minimum server specifications for each node are as follows:


Node	CPU	CPU frequency	vCPUs	Memory	Storage				Number of fixed public IP address
Node	CPU	CPU frequency	vCPUs	Memory	System disk	Data disk	Disk IOPS	Disk throughput	Number of fixed public IP address
Worker-1	x86‑64	2.5 GHz	8	32 GB	150 GB	/	3000	125 MB/s (250MB/s recommended)	1
Worker-2			8	32 GB	150 GB	/			1
Worker-3			8	32 GB	150 GB	/			1
Worker-4			8	32 GB	150 GB	/			1
Worker-5			4	8 GB	150 GB	/			/
Worker-6			4	8 GB	150 GB	/			/
Worker-7			8	16 GB	150 GB	/			/
Worker-8			4	8 GB	50 GB	> 1 TB			/
Worker-9			4	8 GB	50 GB	> 1 TB			/
LB-1		/	2	4 GB	/	/	/	/	1
LB-2		/	2	4 GB	/	/	/	/	1
Total	/	/	60	184 GB	1150 GB	＞ 2 TB	/	/	5

Important: The Data disk on each of Worker-8 and Worker-9 must have sufficient additional disk capacity to store prompts, call recordings, and other data.

Minimum server specifications for 300 concurrent calls

This deployment consists of 11 nodes - 9 worker nodes for Kubernetes Cluster and 2 load balancer nodes for Load Balancer Cluster. Node workloads are as follows.


Cluster	Node	Description
Kubernetes	Worker-1	Host core service components (e.g., Edge, HTTP Proxy, and Database), allowing flexible pod scheduling for efficient resource utilization.
	Worker-2
	Worker-3
	Worker-4
	Worker-5	Dedicated to media processing component to increase processing capacity.
	Worker-6
	Worker-7	Dedicated to monitoring components to isolate monitoring workloads from core call services.
	Worker-8	Dedicated to block storage component to provide persistent storage for prompts, call recordings, and other data.
	Worker-9
Load Balancer	LB-1	Operate in an active-standby mode to ensure high availability through failover. The active LB node performs health checks and distributes traffic.
Load Balancer	LB-2

The minimum server specifications for each node are as follows:


Node	CPU	CPU frequency	vCPUs	Memory	Storage				Number of fixed public IP address
Node	CPU	CPU frequency	vCPUs	Memory	System disk	Data disk	Disk IOPS	Disk throughput	Number of fixed public IP address
Worker-1	x86‑64	2.5 GHz	8	32 GB	150 GB	/	3000	125 MB/s (250MB/s recommended)	1
Worker-2			8	32 GB	150 GB	/			1
Worker-3			8	32 GB	150 GB	/			1
Worker-4			8	32 GB	150 GB	/			1
Worker-5			8	16 GB	150 GB	/			/
Worker-6			8	16 GB	150 GB	/			/
Worker-7			8	16 GB	150 GB	/			/
Worker-8			4	8 GB	50 GB	> 1 TB			/
Worker-9			4	8 GB	50 GB	> 1 TB			/
LB-1		/	2	4 GB	/	/	/	/	1
LB-2		/	2	4 GB	/	/	/	/	1
11	/	/	68	200 GB	1150 GB	＞ 2 TB	/	/	5

Important: The Data disk on each of Worker-8 and Worker-9 must have sufficient additional disk capacity to store prompts, call recordings, and other data.

Minimum server specifications for 500 concurrent calls

This deployment consists of 12 nodes - 10 worker nodes for Kubernetes Cluster and 2 load balancer nodes for Load Balancer Cluster. Node workloads are as follows.


Cluster	Node	Description
Kubernetes	Worker-1	Host core service components (e.g., Edge, HTTP Proxy, and Database), allowing flexible pod scheduling for efficient resource utilization.
	Worker-2
	Worker-3
	Worker-4
	Worker-5	Dedicated to media and SIP processing component to increase capacity of media processing, SIP registration, and other non-call-related task processing.
	Worker-6
	Worker-7
	Worker-8	Dedicated to monitoring components to isolate monitoring workloads from core call services.
	Worker-9	Dedicated to block storage component to provide persistent storage for prompts, call recordings, and other data.
	Worker-10
Load Balancer	LB-1	Operate in an active-standby mode to ensure high availability through failover. The active LB node performs health checks and distributes traffic.
Load Balancer	LB-2

The minimum server specifications for each node are as follows:


Node	CPU	CPU frequency	vCPUs	Memory	Storage				Number of fixed public IP address
Node	CPU	CPU frequency	vCPUs	Memory	System disk	Data disk	Disk IOPS	Disk throughput	Number of fixed public IP address
Worker-1	x86‑64	2.5 GHz	8	32 GB	150 GB	/	3000	125 MB/s (250MB/s recommended)	1
Worker-2			8	32 GB	150 GB	/			1
Worker-3			8	32 GB	150 GB	/			1
Worker-4			8	32 GB	150 GB	/			1
Worker-5			8	16 GB	150 GB	/			/
Worker-6			8	16 GB	150 GB	/			/
Worker-7			8	16 GB	150 GB	/			/
Worker-8			8	16 GB	150 GB	/			/
Worker-9			4	8 GB	50 GB	> 1 TB			/
Worker-10			4	8 GB	50 GB	> 1 TB			/
LB-1		/	2	4 GB	/	/	/	/	1
LB-2		/	2	4 GB	/	/	/	/	1
Total	/	/	76	216 GB	1300 GB	＞ 2 TB	/	/	5

Important: The Data disk on each of Worker-9 and Worker-10 must have sufficient additional disk capacity to store prompts, call recordings, and other data.

Minimum server specifications for 800 concurrent calls

This deployment consists of 16 nodes - 14 worker nodes for Kubernetes Cluster and 2 load balancer nodes for Load Balancer Cluster. Node workloads are as follows.


Cluster	Node	Description
Kubernetes	Worker-1	Host core service components (e.g., Edge, HTTP Proxy, and Database), allowing flexible pod scheduling for efficient resource utilization.
	Worker-2
	Worker-3
	Worker-4
	Worker-5	Dedicated to media and SIP processing component to increase capacity of media processing, SIP registration, and other non-call-related task processing.
	Worker-6
	Worker-7
	Worker-8
	Worker-9	Dedicated to media handling component to raise the handling capacity.
	Worker-10
	Worker-11	Dedicated to monitoring components to isolate monitoring workloads from core call services.
	Worker-12	Dedicated to block storage component to provide persistent storage for prompts, call recordings, and other data.
	Worker-13
	Worker-14
Load Balancer	LB-1	Operate in an active-standby mode to ensure high availability through failover. The active LB node performs health checks and distributes traffic.
Load Balancer	LB-2

The minimum server specifications for each node are as follows:


Node	CPU	CPU frequency	vCPUs	Memory	Storage				Number of fixed public IP address
Node	CPU	CPU frequency	vCPUs	Memory	System disk	Data disk	Disk IOPS	Disk throughput	Number of fixed public IP address
Worker-1	x86‑64	2.5 GHz	8	32 GB	150 GB	/	3000	125 MB/s (250MB/s recommended)	1
Worker-2			8	32 GB	150 GB	/			1
Worker-3			8	32 GB	150 GB	/			1
Worker-4			8	32 GB	150 GB	/			1
Worker-5			8	16 GB	150 GB	/			/
Worker-6			8	16 GB	150 GB	/			/
Worker-7			8	16 GB	150 GB	/			/
Worker-8			8	32 GB	150 GB	/			/
Worker-9			4	8 GB	150 GB	/			1
Worker-10			4	8 GB	150 GB	/			1
Worker-11			8	16 GB	150 GB	/			/
Worker-12			4	8 GB	50 GB	> 2 TB			/
Worker-13			4	8 GB	50 GB	> 2 TB			/
Worker-14			4	8 GB	50 GB	> 2 TB			/
LB-1		/	2	4 GB	/	/	/	/	1
LB-2		/	2	4 GB	/	/	/	/	1
Total	/	/	96	256 GB	1800 GB	＞ 6 TB	/	/	7

Important: The Data disk on each of Worker-12, Worker-13 and Worker-14 must have sufficient additional disk capacity to store prompts, call recordings, and other data.

Minimum server specifications for 1000 concurrent calls

This deployment consists of 17 nodes - 15 worker nodes for Kubernetes Cluster and 2 load balancer nodes for Load Balancer Cluster. Node workloads are as follows.


Cluster	Node	Description
Kubernetes	Worker-1	Host core service components (e.g., Edge, HTTP Proxy, and Database), allowing flexible pod scheduling for efficient resource utilization.
	Worker-2
	Worker-3
	Worker-4
	Worker-5	Dedicated to media and SIP processing component to increase capacity of media processing, SIP registration, and other non-call-related task processing.
	Worker-6
	Worker-7
	Worker-8	Dedicated to media handling component to raise the handling capacity.
	Worker-9
	Worker-10	Dedicated to reporting component to isolate resource archiving for CDR, recordings, voicemail and other data.
	Worker-11
	Worker-12	Dedicated to monitoring components to isolate monitoring workloads from core call services.
	Worker-13	Dedicated to block storage component to provide persistent storage for prompts, call recordings, and other data.
	Worker-14
	Worker-15
Load Balancer	LB-1	Operate in an active-standby mode to ensure high availability through failover. The active LB node performs health checks and distributes traffic.
Load Balancer	LB-2

The minimum server specifications for each node are as follows:


	CPU	CPU frequency	vCPUs	Memory	Storage				Number of fixed public IP address
	CPU	CPU frequency	vCPUs	Memory	System disk	Data disk	Disk IOPS	Disk throughput	Number of fixed public IP address
Worker-1	x86‑64	2.5 GHz	8	32 GB	150 GB	/	3000	125 MB/s (250MB/s recommended)	1
Worker-2			8	32 GB	150 GB	/			1
Worker-3			8	32 GB	150 GB	/			1
Worker-4			8	32 GB	150 GB	/			1
Worker-5			8	32 GB	150 GB	/			/
Worker-6			8	32 GB	150 GB	/			/
Worker-7			8	32 GB	150 GB	/			/
Worker-8			8	8 GB	150 GB	/			1
Worker-9			8	8 GB	150 GB	/			1
Worker-10			8	8 GB	150 GB	/			/
Worker-11			8	8 GB	150 GB	/			/
Worker-12			8	32 GB	150 GB	/			/
Worker-13			8	8 GB	50 GB	> 2 TB			/
Worker-14			8	8 GB	50 GB	> 2 TB			/
Worker-15			8	8 GB	50 GB	> 2 TB			/
LB-1		/	2	4 GB	/	/	/	/	1
LB-2		/	2	4 GB	/	/	/	/	1
Total	/	/	112	288 GB	1950 GB	＞ 6 TB	/	/	7

Important: The Data disk on each of Worker-13, Worker-14 and Worker-15 must have sufficient additional disk capacity to store prompts, call recordings, and other data.

Additional disk capacity

For each node dedicated to running block storage component, the Data disk must provide additional storage capacity for prompts, call recordings, and other data.

To estimate your storage requirements based on your workload, refer to the example below or contact Yeastar Support for further recommendations.

Example

A 3-minute stereo call recording requires approximately 6 MB of storage. At 48000 calls per day, call recordings require approximately 105 TB per year (365 days). To retain one year of recordings, allocate 105 TB of storage to each node (dedicated to Longhorn), plus an additional 25% buffer for other data, as needed.

Operating system

The recommended operating system are listed below. To deploy NovoOne on other operating systems, contact Yeastar Support for further recommendations.


Platform	Distribution
Linux	Ubuntu 24.04 LTS - 64 bit

Network

Important: The private subnet of each node MUST be set to one of the following subnet options, otherwise the node deployment would fail.

10.0.0.0/8
172.16.0.0/12
192.168.0.0/16

Ethernet port

Each node needs 1 Ethernet port.

Bandwidth


Item	Minimum Upstream/Downstream Bandwidth
Internal Network Bandwidth	10 Gbps
External Network Bandwidth	250 Mbps

Port

Inter-Node Communication

All Kubernetes worker nodes must be able to communicate with each other over all ports and protocols.
All load balancer nodes must be able to communicate over all ports and protocols with Kubernetes worker nodes configured with fixed public IP addresses.

Public Network Access

For Kubernetes worker nodes configured with fixed public IP addresses, open the following ports on the firewall:


Protocol	Port
TCP	22
TCP, UDP	5090
TCP	5091
TCP, UDP	30000-59999

For load balancer nodes, open the following ports on the firewall:


Protocol	Port
TCP	22
TCP	80
TCP, UDP	443
TCP, UDP	3478
TCP, UDP	5060
TCP	5061
TCP, UDP	5092
TCP	5093