Data Transfer Slowness Caused by Sub-Partition Hard Limits and System Global Link Configuration
search cancel

Data Transfer Slowness Caused by Sub-Partition Hard Limits and System Global Link Configuration

book

Article ID: 449523

calendar_today

Updated On:

Products

PS-S500

Issue/Introduction

Automated high-throughput data transfers between cloud instances and on-premises infrastructure experience severe performance degradation (taking up to 4x to 5x longer) when global Packet Shaping is enabled.

Cause

Specific sub-partitions housing cloud traffic are configured with low maximum bandwidth ceilings (such as 20.0M* or 7.0M). When shaping is active, the system strictly caps traffic to these limits.

Resolution

To prevent a single cloud class from taking over the shared uncommitted bandwidth pool, use ipfilter passthrough. This bypasses classification and shaping at the network interface driver level during data transfers, allowing raw wire speed without modifying the sub-partition tree.

ipfilter passthrough <device> src|dst <ipaddress>[:<port>] [<mask>]
[prot:tcp|udp|icmp]