HP t505 Technology Overview - Page 8

Technology Overview | HP Velocity LiveTCP latency mitigation Real-world IP networks introduce latency and packet loss to application flows. A primary cause of these issues is network congestion. Latency and packet loss degrade the end-user QoE, especially for applications that use TCP for reliable data transmission, such as streaming video, chat, remote desktop, and file transfers. This results in unresponsive user interfaces for remote desktop users, extended wait times for file transfers, and choppy video with reduced frame rates. TCP congestion avoidance mechanisms TCP is designed to send data as rapidly as possible until it detects packet loss. Once packet loss is detected, TCP halts data delivery and reduces transmission speed while lost packets are retransmitted. Built-in algorithms enable TCP to avoid excessive congestion of the network. As packet loss dissipates, TCP increases the speed of data delivery. TCP algorithms TCP can be controlled by different algorithms to accommodate different types of networks. However, only one network type at a time can be accommodated under native TCP. For example, LANs are characterized by low packet loss and high bandwidth. Remote office networks typically have high packet loss and high latency. Wi-Fi networks have high jitter (that is, the latency is highly variable) and variable throughput capacity. TCP tuning Most operating systems allow only a single TCP algorithm to be active with a single set of settings. If a network is optimized for LAN traffic, WAN connections are less efficient. While the TCP stack can be tuned, the effort required by network administrators is a deterrent. LiveTCP automatically provides tuning, which boosts TCP throughput over all networks and provides latency mitigation for streaming applications. LiveTCP congestion control LiveTCP adapts its algorithm to automatically provide congestion control that is optimized independently for each TCP flow (LAN, remote network, Wi-Fi) and the conditions in the network. LiveTCP achieves this without requiring intervention from IT. LiveTCP improves on native TCP by accelerating the speed at which thin-client protocols transmit data. For example, for RDP, which is highly sensitive to latency, LiveTCP can improve transmission speeds up to 10 times; for HP RGS, two to three times; for VMware Blast, up to two times; and for ICA, two times. In public networks where congestion is often a problem, such as at airports, hotels, and coffee shops, LiveTCP improves overall throughput by more effectively managing shared bandwidth. LiveTCP is configured to provide latency mitigation for application flows between two HP Velocity-enabled endpoints (that is, in a peering configuration) or from an HP Velocity-enabled endpoint to a non HP Velocity-enabled endpoint (single-sided). In both configurations (Figure 4 and Figure 5), the optimized LiveTCP algorithm takes precedence over the native TCP algorithm. Figure 4: LiveTCP algorithm in a peering configuration Figure 5: LiveTCP algorithm in a single-sided configuration 6