Specifications of Dedicated Load Balancers

When you create a dedicated load balancer, you can select elastic or fixed specifications based on your service requirements. Table 1 lists the differences between the two specifications.

**Table 1** Specifications comparison
Item	Elastic	Fixed
Application scenarios	For fluctuating traffic When you need to use resources temporarily and for urgent purposes	For stable traffic When you need to use resources for a long term
Network (TCP/UDP/TLS) load balancer performance	The performance multiplies as the number of AZs increases. Table 3 shows the maximum performance in an AZ.	The performance multiplies as the number of AZs increases. Table 6 shows the maximum performance in an AZ.
Application (HTTP/HTTPS) load balancer performance	The performance multiplies as the number of AZs increases. Table 3 shows the maximum performance in an AZ.	The performance multiplies as the number of AZs increases. Table 7 shows the maximum performance in an AZ.
Billing mode	Pay-per-use	Pay-per-use Yearly/Monthly
Billing items	LCU Load balancer	LCU
Capabilities	Same

In the private network access scenario, if clients are in the same AZ as the load balancer, requests are distributed by the load balancer in this AZ. If the load balancer goes down, requests are distributed by the load balancer in another AZ.

If the load balancer is healthy but the connections that the load balancer needs to handle exceed the amount defined in the specifications, service may be interrupted. To address this issue, you need to upgrade specifications.

You can view the monitoring metrics supported by AZ to check whether private network traffic has exceeded the upper limit.

Elastic Specifications

If your service traffic fluctuates greatly, you can choose elastic specifications and select network load balancing (TCP/UDP/TLS) or application load balancing (HTTP/HTTPS), or both that best meet your service needs.

The listener protocol must match the load balancing type. For example, if you select application load balancing (HTTP/HTTPS), you can only add an HTTP or HTTPS listener to this load balancer.

Table 2 describes the dimensions about elastic specifications. When your traffic exceeds the specifications defined in Table 3, new requests will be discarded.

**Table 2** Elastic specification dimensions
Maximum Connections	Indicates the maximum number of concurrent connections that a load balancer can handle per minute. If the number reaches the maximum connections that is defined in the elastic specifications, new requests will be discarded to ensure the performance of established connections.
Connection Per Second (CPS)	Indicates the number of new connections that a load balancer can establish per second. If the number reaches the CPS that is defined in the elastic specifications, new requests will be discarded to ensure the performance of established connections.
Query Per Second (QPS)	Indicates the number of HTTP or HTTPS requests sent to a backend server per second. If the QPS reaches that is defined in the elastic specifications, new requests will be discarded to ensure the performance of established connections.
Bandwidth (Mbit/s)	Indicates the maximum amount of data that can be transmitted over a connection per second.

**Table 3** Maximum elastic specifications
Protocol	Maximum Connections	CPS	QPS	Bandwidth (Mbit/s)
Network load balancing (TCP/UDP)	20,000,000	400,000	N/A	10,000
Network load balancing (TLS)	20,000,000	400,000	N/A	10,000
Application load balancing (HTTP)	8,000,000	80,000	160,000	10,000
Application load balancing (HTTPS)	8,000,000	80,000	160,000	10,000

Available elastic specifications are displayed on the console and may vary depending on regions.

Fixed Specifications

Load balancers are available in different fixed specifications. Choose the specifications that best meet your needs. When your traffic exceeds what defined in your selected specifications, new requests will be discarded. Each specification has the following dimensions.

**Table 4** Fixed specification dimensions
Maximum Connections	Indicates the maximum number of concurrent connections that a load balancer can handle per minute. If the number reaches the maximum connections that is defined in your selected fixed specifications, new requests will be discarded to ensure the performance of existing connections.
CPS	Indicates the number of new connections that a load balancer can establish per second. If the number reaches the CPS that is defined in your selected fixed specifications, new requests will be discarded to ensure the performance of established connections. HTTPS listeners need to create SSL handshakes to establish connections with clients, and such SSL handshakes occupy more system resources than HTTP listeners. For example, a small I application load balancer can establish 2,000 new HTTP connections per second but only 200 new HTTPS connections per second. For details, see Table 5.
QPS	Indicates the number of HTTP or HTTPS requests sent to a backend server per second. If the QPS reaches that is defined in your selected fixed specifications, new requests will be discarded to ensure the performance of established connections.
Bandwidth (Mbit/s)	Indicates the maximum amount of data that can be transmitted over a connection per second.

For a small I application load balancer:

If you only add an HTTP listener, the load balancer can establish up to 2,000 new HTTP connections.
If you only add an HTTPS listener, the load balancer can establish up to 200 new HTTPS connections.

If you add an HTTPS listener and an HTTP listener, the new connections are calculated using the following formula:

New connections = New HTTP connections + New HTTPS connections x Ratio of HTTP connections to HTTPS connections

For a small I application load balancer, the ratio of HTTP connections to HTTPS connections is 10. For details, see Table 5.

**Table 5** New connections that a small I application load balancer can establish
Parameter	Scenario 1	Scenario 2
New HTTP connections	1,000	1,000
New HTTPS connections	50	150
New HTTP and HTTPS connections	1,000 + 50 x 10 = 1,500	1,000 + 150 x 10 = 2,500
Description	The new connections do not reach the CPS (HTTP) that a small I application load balancer can handle, so new requests can be properly routed.	The new connections exceed the CPS (HTTP) that a small I application load balancer can handle, so new requests will be discarded.

Details in Table 5 are for reference only.

Table 6 and Table 7 list the fixed specifications of dedicated load balancers.

Available fixed specifications are displayed on the console and may vary depending on the resources in different regions.
The listener protocol must match the load balancing type. For example, if you create an application load balancer, you can only add an HTTP or HTTPS listener to this load balancer.

**Table 6** Fixed specifications for a network load balancer (TCP/UDP/TLS)
Type	Maximum Connections (TCP/UDP)	Maximum Connections (TLS)	CPS (TCP/UDP)	CPS (TLS)	Bandwidth (Mbit/s)	LCUs in an AZ
Small I	500,000	30,000	10,000	500	50	10
Small II	1,000,000	60,000	20,000	1,000	100	20
Medium I	2,000,000	120,000	40,000	2,000	200	40
Medium II	4,000,000	240,000	80,000	4,000	400	80
Large I	10,000,000	600,000	200,000	10,000	1,000	200
Large II	20,000,000	1,200,000	400,000	20,000	2,000	400

**Table 7** Fixed specifications for an application load balancer (HTTP/HTTPS)
Type	Maximum Connections	CPS (HTTP)	CPS (HTTPS)	QPS (HTTP)	QPS (HTTPS)	Bandwidth (Mbit/s)	LCUs in an AZ
Small I	200,000	2,000	200	4,000	2,000	50	10
Small II	400,000	4,000	400	8,000	4,000	100	20
Medium I	800,000	8,000	800	16,000	8,000	200	40
Medium II	2,000,000	20,000	2,000	40,000	20,000	400	100
Large I	4,000,000	40,000	4,000	80,000	40,000	1,000	200
Large II	8,000,000	80,000	8,000	160,000	80,000	2,000	400

If you add multiple listeners to a load balancer, the sum of QPS values of all listeners cannot exceed the QPS defined in each specification.
The bandwidth is the upper limit of the inbound or the outbound traffic. For example, for small I load balancers, the inbound or outbound traffic cannot exceed 50 Mbit/s.
The bandwidth included in each specification is the maximum bandwidth provided by ELB. If the maximum bandwidth is exceeded, the network performance may be affected.

Previous topic: Network Traffic Paths

Next topic: Notes and Constraints