QoS Mechanisms: DiffServ and IntServ

IntServ: Explicit, Per-Flow Reservations

IntServ (Integrated Services) works by having an application explicitly request a guaranteed amount of bandwidth and quality for its specific traffic flow, before that traffic even starts - every router along the path must agree to reserve those exact resources for that specific flow.

Application: "I need a guaranteed 2 Mbps, low-latency path, for this specific video call"
Every router along the path: "Confirmed, resources reserved specifically for this flow"

Why IntServ Struggles to Scale

Requiring every router along a path to track and maintain a reservation for every single individual flow becomes an enormous administrative and resource burden on a large, busy network - this scalability problem is exactly why IntServ, while conceptually clean, saw limited widespread real-world adoption.

DiffServ: A Simpler, Class-Based Approach

DiffServ (Differentiated Services) takes a different approach: rather than reserving resources per individual flow, it simply marks each packet with a priority class using a field in its header, and routers make forwarding decisions based on that marked class - no per-flow reservation or tracking required.

Packet marked "Expedited Forwarding" (high priority)  -> routers handle it preferentially
Packet marked "Best Effort" (default, lowest priority)  -> handled normally, no special treatment

Why DiffServ Scales Far Better

Routers only need to look at a packet's marked class - a simple, fixed, small number of categories - rather than tracking potentially millions of individual flows. This dramatically lower overhead is exactly why DiffServ became the practical, dominant approach used across most real-world networks today.

Comparing the Two