As part of workspace digitization, a large enterprise has migrated all their users to Desktop as a Sen/ice (DaaS), by hosting the backend system in their on-premises data center. Some of the branches

Question

Options

• Atail drop
• Btraffic shaping
• CWRED
• Dtraffic policing

• A
  10% (4)
• B
  5% (2)
• C
  67% (28)
• D
  19% (8)

Explanation

Branch users connect to the on-premises DaaS backend over WAN links, which are subject to congestion. Local and head office users are on the LAN and avoid this bottleneck. The periodic disconnections are symptomatic of WAN queue congestion causing TCP sessions to reset or time out. WRED (Weighted Random Early Detection) is a congestion-avoidance mechanism that proactively drops lower-priority packets as queue depth increases - before the queue fills completely. This prevents TCP global synchronization (where all TCP flows simultaneously back off), keeps latency predictable, and maintains stable interactive sessions like those used by DaaS. Tail drop (A) reactively drops all packets only when the queue is full, worsening global sync. Traffic shaping (B) buffers and smooths egress traffic but does not prioritize or protect sensitive flows. Traffic policing (D) drops or re-marks excess traffic at a hard rate limit, which can actively cause the very disconnections being reported.

No community discussion yet for this question.