Network Performance
Latency floors, bandwidth-delay product, and what no code change fixes.
5 to work through
-
intermediate
A managed database platform sees connection exhaustion when customers run serverless functions against it. Why does this happen and what is the fix?
2 min answer -
advanced
A connected-vehicle fleet sends telemetry over unreliable mobile networks with limited bandwidth and intermittent connectivity. What protocol and session characteristics matter, and what is the wrong choice?
2 min answer -
advanced
A global SaaS product has a 200 ms p95 API budget. Users in Sydney see 900 ms against a single US-East deployment. The application team says the service responds in 40 ms. Who is right and what do you do?
2 min answer -
advanced
A search tier fans every query out to 60 shards and merges the replies. The merged p99 jumps from 40 ms to 240 ms, with a few requests at 440 ms. Shard CPU is flat, no individual shard's p99 moved, and the latency increments are almost exactly 200 ms. What failed?
3 min answer -
advanced
A short-video platform finds that video startup time, not playback quality, is the strongest predictor of engagement. Which network-level optimisations matter most?
2 min answer
5 terms in this topic
Bandwidth-Delay Product
Bandwidth multiplied by round-trip time — the amount of data that must be in flight to keep a link fully utilised.
conceptConnection Fan-In
The problem created when many short-lived compute instances each require a connection to a stateful dependency that can only support a small number -…
conceptECMP Flow Pinning
Because a router picks one of several equal-cost paths by hashing a flow's five-tuple, one connection is confined to one link, so aggregate capacity …
conceptNetwork Performance
Latency is bounded by physics and bandwidth is bought — so architectural performance work is mostly about reducing round trips and moving data closer.
metricRound-Trip Time
The time for a packet to travel to a destination and back — a physical floor that no application optimisation can reduce.
Neighbouring topics
Networking
General material on the network path underneath an architecture.
TCP/IP
Handshakes, congestion control, head-of-line blocking and connection reuse.
HTTP/1.1, HTTP/2 & HTTP/3
Multiplexing, prioritisation and the transport each is built on.
DNS
Resolution, TTL behaviour, traffic steering and failover latency.
TLS & Certificates
Encryption, integrity, authentication, termination points and expiry.
Layer 4 vs Layer 7
Connection-level versus request-level balancing, and what each unlocks.
Reverse Proxies
One place for TLS, routing, caching, compression and rate limiting.
Content Delivery Networks
Edge caching, origin offload, spike absorption and dynamic content.
Firewalls & Security Groups
Default-deny, stateful rules, and restricting egress as well as ingress.
NAT & Egress
Outbound-only connectivity, its per-gigabyte cost and its zone binding.
VPC Design
Address planning, peering and the ranges you can never resize.
Subnetting
Tiering, zone binding, and sizing for an address-hungry platform.
Routing & BGP
How traffic finds you, anycast, and route withdrawal as a failure mode.
Service Mesh Networking
Sidecar proxies supplying mTLS, retries and traffic policy.
WebSockets & Realtime
Persistent bidirectional connections and the capacity model they impose.
gRPC Transport
HTTP/2 multiplexing, binary encoding and streaming semantics.
API Gateways
The single entry point, and the business logic that must stay out of it.
Private Connectivity
Private endpoints, peering and dedicated links to managed services.
Network Troubleshooting
Flow logs, packet paths, and localising a problem to a hop.