[Forge][Chaos] remove jitter, make inter-region BW 300 Mbps (aptos-la…

…bs#10277)

### Description

The changes are based on observations while measuring network performance and reading more into the dataset used.
* Jitter: My understanding is that re-ordering of packets should be pretty rare in the real world, while the previous jitter configs would introduce re-ordering quite frequently. Unless we have a strong belief that jitter is present in our networks, we shouldn't mess with this.
* Inter-region bitrate: The numbers this was based on are iperf with a single TCP stream. The results correlate strongly with RTT, which suggests that RTT is the limiting factor, so there's no real reason to constrain BW itself. For now, 300 Mbps is as fast as our network stack will go for 100+ ms RTT.

testsuite/testcases/src/data/four_region_link_stats.csv

@@ -1,13 +1,13 @@
 sending_region,receiving_region,bitrate_bps,avgrtt
-gcp--us-central1,aws--eu-west-1,86900736,103.435
-gcp--us-central1,aws--ap-northeast-1,65798144,133.996
-gcp--us-central1,aws--sa-east-1,57933824,145.483
-aws--sa-east-1,gcp--us-central1,200146944,145.703
-aws--sa-east-1,aws--eu-west-1,48365568,176.894
-aws--sa-east-1,aws--ap-northeast-1,32505856,255.289
-aws--eu-west-1,gcp--us-central1,275513344,104.169
-aws--eu-west-1,aws--sa-east-1,46530560,176.813
-aws--eu-west-1,aws--ap-northeast-1,40632320,198.555
-aws--ap-northeast-1,gcp--us-central1,214827008,128.999
-aws--ap-northeast-1,aws--eu-west-1,41287680,198.539
-aws--ap-northeast-1,aws--sa-east-1,32243712,255.323
+gcp--us-central1,aws--eu-west-1,300000000,103.435
+gcp--us-central1,aws--ap-northeast-1,300000000,133.996
+gcp--us-central1,aws--sa-east-1,300000000,145.483
+aws--sa-east-1,gcp--us-central1,300000000,145.703
+aws--sa-east-1,aws--eu-west-1,300000000,176.894
+aws--sa-east-1,aws--ap-northeast-1,300000000,255.289
+aws--eu-west-1,gcp--us-central1,300000000,104.169
+aws--eu-west-1,aws--sa-east-1,300000000,176.813
+aws--eu-west-1,aws--ap-northeast-1,300000000,198.555
+aws--ap-northeast-1,gcp--us-central1,300000000,128.999
+aws--ap-northeast-1,aws--eu-west-1,300000000,198.539
+aws--ap-northeast-1,aws--sa-east-1,300000000,255.323

testsuite/testcases/src/data/two_region_link_stats.csv

@@ -1,3 +1,3 @@
 sending_region,receiving_region,bitrate_bps,avgrtt
-aws--sa-east-1,aws--ap-northeast-1,32505856,255.289
-aws--ap-northeast-1,aws--sa-east-1,32243712,255.323
+aws--sa-east-1,aws--ap-northeast-1,300000000,255.289
+aws--ap-northeast-1,aws--sa-east-1,300000000,255.323

testsuite/testcases/src/multi_region_network_test.rs

@@ -10,7 +10,10 @@ use std::collections::BTreeMap;
 
 /// The link stats are obtained from https://github.com/doitintl/intercloud-throughput/blob/master/results_202202/results.csv
 /// The four regions were hand-picked from the dataset to simulate a multi-region setup
-/// with high latencies and low bandwidth.
+/// with high latencies.
+/// Note, we restrict bandwidth to 300 Mbps between all regions. The reasoning is that the dataset
+/// is measuring TCP bandwidth only which is primarily affected by RTT, and not the actual bandwidth
+/// across the regions, which would vary according to competing traffic, etc.
 const FOUR_REGION_LINK_STATS: &[u8] = include_bytes!("data/four_region_link_stats.csv");
 /// The two regions were chosen as the most distant regions among the four regions set.
 const TWO_REGION_LINK_STATS: &[u8] = include_bytes!("data/two_region_link_stats.csv");
@@ -103,7 +106,7 @@ pub struct InterRegionNetEmConfig {
 impl Default for InterRegionNetEmConfig {
     fn default() -> Self {
         Self {
-            delay_jitter_ms: 20,
+            delay_jitter_ms: 0,
             delay_correlation_percentage: 50,
             loss_percentage: 3,
             loss_correlation_percentage: 50,
@@ -158,7 +161,7 @@ impl Default for IntraRegionNetEmConfig {
         Self {
             bandwidth_rate_mbps: 10 * 1000, // 10 Gbps
             delay_latency_ms: 50,
-            delay_jitter_ms: 5,
+            delay_jitter_ms: 0,
             delay_correlation_percentage: 50,
             loss_percentage: 1,
             loss_correlation_percentage: 50,