D-Link DFL-2500 - Manual (Page 284)

If Connection Rate is applied instead, R1 and R2 will be sent to the same server because of
stickiness, but the subsequent requests R3 and R4 will be routed to another server since the number
of new connections on each server within the Window Time span is counted in for the distribution.
Figure 10.8. Stickiness and Connection Rate
Regardless which algorithm is chosen, if a server goes down, traffic will be sent to other servers.
And when the server comes back online, it can automatically be placed back into the server farm
and start getting requests again.
10.3.5. Server Health Monitoring
SLB uses Server Health Monitoring to continuously check the condition of the servers in an SLB
configuration. SLB monitors different OSI layers to check the connection rate for each server as
well as its current state. Regardless of the algorithm, if a server fails, SLB will not send any more
requests until it the server recovers to full functionality.
SLB will use the default routing table unless the administrator sets a specific routing table location.
D-Link Server Load Balancing provides the following monitoring modes:
ICMP Ping This works at OSI layer 3. SLB will ping the IP address of each individual
server in the server farm. This will detect any failed servers.
TCP Connection This works at OSI layer 4. SLB attempts to connect to a specified port on
each server. For example, if a server is specified as running web services on
port 80, the SLB will send a TCP SYN request to that port. If SLB does not
receive a TCP SYN/ACK back, it will mark port 80 on that server as down.
SLB recognizes the conditions no response, normal response or closed port
response from servers.
10.3.6. SLB_SAT Rules
10.3.5. Server Health Monitoring Chapter 10. Traffic Management
284