What are the Load Balancing Algorithms on TR7?
About
What Load Balancing Algorithms Are Available on TR7?
There are 16 different types of load balancing algorithms available on the TR7 ASP. Requests coming to the vService are distributed to the organization's services according to the selected load balancing algorithms. This way, traffic is distributed to the redundant servers based on the algorithms set according to your needs, helping to lighten the server's workload. The load balancing algorithms available on the TR7 ASP are listed below.
- Round Robin
- Weighted Round Robin
- Least Connection
- First
- Random
- Fastest
- Fastest+
- Source (Self-Persistent)
- URI (Self-Persistent)
- URL Param (Self-Persistent)
- HDR (Self-Persistent)
- RDP Cookie (Self-Persistent)
- Locality-Based Least-Connection
- Weighted Least-Connection
- Source Hash
- Destination Hash
Interface
vService Type > HTTP
Round Robin
The load is distributed regardless of the weight ratios of the backend services.
Weighted Round Robin
The load is distributed based on the weight ratios of the backend services.
Least Connection
Backend services with the least number of active connections at any given moment are prioritized according to their weight ratios. This is recommended for services that require long sessions, such as LDAP or SQL.
First
The load is distributed to the first available backend service until its maximum connection limit is reached.
Random
It is used for random selection of backend services, taking into account the weight value.
Fastest
The load is distributed to the most suitable backend service based on the selected parameter.
- Least Response Time
- Least Connection Time
- Least Queue Time
- Least Connection
- Least Queues
- Least Connection Error
- Least Aborted Connections
- Least Used Connections
Fastest+
The load is distributed to the most suitable backend service based on the first selected parameter. In case of a tie based on the first parameter, the distribution is made according to the second parameter.
Fastest Opt-1
- Least Response Time
- Least Connection Time
- Least Queue Time
- Least Connection
- Least Queues
- Least Connection Error
- Least Aborted Connections
- Least Used Connections
Fastest Opt-2
- Least Response Time
- Least Connection Time
- Least Queue Time
- Least Connection
- Least Queues
- Least Connection Error
- Least Aborted Connections
- Least Used Connections
Source (Self-Persistent)
The hash value of the source IP address is calculated and distributed according to the total weight of the backend services. The same client is always directed to the same backend service.
URI (Self-Persistent)
The hash value of the URI, up to the portion before the question mark (?), is calculated and distributed according to the total weight of the backend services. The same URI is always directed to the same backend service.
URL Param (Self-Persistent)
It is used to track the User ID (User ID Identifier). The same User ID is always directed to the same backend service.
HDR (Self-Persistent)
The specified HTTP header name is searched within each HTTP request. The same header contents are always directed to the same backend service.
Interface
vService Type > L7 TCP
Round Robin
The load is distributed regardless of the weight ratios of the backend services.
Weighted Round Robin
The load is distributed based on the weight ratios of the backend services.
Least Connection
Backend services with the fewest active connections at any given moment are prioritized according to their weight ratios. This is recommended for services that require long sessions, such as LDAP or SQL.
First
The load is distributed to the first available backend service until its maximum connection limit is reached.
Random
It is used for random selection of backend services, taking into account their weight values.
Fastest
The load is distributed to the most suitable backend service based on the selected parameter.
- Least Response Time
- Least Connection Time
- Least Queue Time
- Least Connection
- Least Queues
- Least Connection Error
- Least Aborted Connections
- Least Used Connections
Fastest+
The load is distributed to the most suitable backend service based on the first selected parameter. In case of a tie based on the first parameter, the distribution is made according to the second parameter.
Fastest Opt-1
- Least Response Time
- Least Connection Time
- Least Queue Time
- Least Connection
- Least Queues
- Least Connection Error
- Least Aborted Connections
- Least Used Connections
Fastest Opt-2
- Least Response Time
- Least Connection Time
- Least Queue Time
- Least Connection
- Least Queues
- Least Connection Error
- Least Aborted Connections
- Least Used Connections
RDP Cookie (Self-Persistent)
The RDP cookie within each incoming TCP packet is identified, and its hash value is calculated. The same user or session ID is always directed to the same backend service.
Interface
vService Type > Network (TCP-L4 and UDP-L4)
Round Robin
The load is distributed regardless of the weight ratios of the backend services.
Weighted Round Robin
The load is distributed based on the weight ratios of the backend services.
Least Connection
Backend services with the fewest active connections at any given moment are prioritized according to their weight ratios. This is recommended for services that require long sessions, such as LDAP or SQL.
Locality-Based Least-Connection
As long as there is no server overload, client traffic from the same IP is directed to the same server. In the case of overload, the traffic is distributed to the server with the fewest connections.
Weighted Least-Connection
The load is distributed to the pool with the fewest connections, based on the server weight.
Source Hash
The load is distributed based on the hash value of the source IP.
Destination Hash
The load is distributed based on the hash value of the destination IP.
Interface
vService Type > Network (ANY-L3)
Round Robin
The load is distributed according to the Round-Robin algorithm.
Weighted Round Robin
The load is distributed according to the Round-Robin algorithm, taking the server weight into account.