Last week, we have largely addressed the TCP profiles and how they help with various application deployment scenarios (/ blogs / 2012/03/22 / air -netscaler-tcp-battery-suit your needs /). We went into the details of embedded profiles, and how you can use them in cases of joint use.
In the case of advances using NetScaler allows you to define your own custom TCP profile. Before you start creating your own profile, you must understand the basic optimization settings that will help you build the profile. After preview shows you the different settings that together define the TCP profile
Come deep into various parameters here :.
Window Scaling
TCP window scaling increases the receiving TCP window size beyond 65535 bytes. It helps to improve overall performance of TCP especially in a high bandwidth and long delay networks. It helps to reduce latency and improve response time over TCP.
Maximum Burst Limit
This parameter controls the burst of packets on the form of thread NetScaler in one attempt. High Limit max burst ensures faster delivery of data in the network without congestion. Packet Burst Limit helps avoid congestion at the link and the intermediate nodes.
Initial Congestion Window Size
TCP initial size of the congestion window determines the number of bytes that can be outstanding at the beginning of the transaction. It allows NetScaler to send these many bytes without worrying about congestion on the wire. The default size is kept as 4 which is 4 * MSS.
TCP Delayed ACK timeout
To avoid sending too much ACK packets on the wire NetScaler implements the delayed ACK mechanism. NetScaler sends delayed ACK with default delay of 0ms. So we accumulate data packets and send ACK to the sending party that if we receive two packets of data in the extension or the timer expires.
size of the packet queue Maximum OOO
maintains TCP Out Of queue commands to keep OOO packets in TCP communication. This system memory setting impact if the size of the queue is long that packets should be stored in the execution memory. Thus, it should be maintained at optimized level depending on the type of network and application characteristics.
Max Packet MSS and MSS
With the MSS setting you can specify TCP maximum segment Size to use for transactions. Default NetScaler 8 sizes retains MSS to be used for TCP transactions: 1460, 1440, 1360, 1212, 956, 536, 384 and 128. Based on what you set here or immediately below is selected for communication. maximum packet MSS is used as a factor to control congestion on the base packages.
Max Packet Forwarding
This allows NetScaler to control the number of packets to be reissued in an attempt. When NetScaler receives a partial ACK and it must broadcast this parameter comes in. This does not affect the RTO retransmissions based.
minimum RTO
The TCP retransmission timeout is calculated each received ACK based on the logic of internal implementation. The default retransmission timeout occurs at 1 second to start and this can be changed with this parameter. For the second RTO retransmission of packets is calculated by No. 2 and No. 4 No. 8 ... ... continues until the last retransmission attempt.
Slow Start Increment
TCP slow start is used to control congestion starting at low rates. The rate is increased exponentially until the first packet loss is noticed. It affects the growth of the congestion window and 2 of default setting will increase the size of the 2 * MSS congestion window by receiving each ACK. It contributes to a faster growth of the congestion window and enables faster data transmission rate. You can increase this parameter if you see the network is sufficiently able to handle the high rate of packets.
TCP buffer size
TCP buffer size is the size of the receive buffer on NetScaler. This buffer size is advertised to clients and servers NetScaler and it controls the ability to send data to NetScaler. The default buffer size is 8K and in most cases it will be safe to increment this while talking to the batteries of internal servers. The size of the buffer is also the impact of the actual application layer in NetScaler as for the case of SSL end is set to 40K and compression, it is set to 96K.
SACK (selective acknowledgment of receipt)
TCP SACK addresses the multiple packet loss problem that reduces the overall capacity of flow. With the selective recognition the receiver can inform the sender about all segments that are successfully received, allowing the sender to retransmit only the segments that have been lost. This technique allows NetScaler improve overall throughput and reduce connection latency.
Nagle algorithm
Nagle algorithm fights with the little packets of problem in the transmission of TCP. Applications such as Telnet and other real-time engines that require each key stroke to be passed to the other side often create very small packages. With the Nagle algorithm NetScaler can buffer these small packets and sends them together to increase the efficiency of the wire. This algorithm needs to work with other TCP optimization techniques in NetScaler.
immediate ACK PUSH
Default NetScaler uses the delayed ACK technique to ensure there are not too many ACK just on the wire bundles. But sometimes applications can wait NetScaler to send ACK immediately without any delay and these packages are often marked by PUSH flag. If we allow this setting NetScaler then responds with ACK to the sender immediately on receiving a packet with PUSH.
Many interesting optimization techniques exposed if the TCP profiles. You have to understand them and their impact on your network deployment / application. One important thing to note is these optimization settings must work together and thus have a certain amount of cross impact as well. Have appropriate profiles can change the way your application behaves and could create a much greater impact on the client side ...
0 Komentar