MVTX2602
Data Sheet
32
Zarlink Semiconductor Inc.
of bandwidth or latency assurances per class. Sometimes it may even reflect an estimate of the traffic mix offered to
the switch. As an added bonus, although we do not assume anything about the arrival pattern, if the incoming traffic
is policed or shaped, we may be able to provide additional assurances about our switch’s performance.
Table 6 shows examples of QoS applications with three transmission priorities, but best effort (P0) traffic may form
a fourth class with no bandwidth or latency assurances.
A class is capable of offering traffic that exceeds the contracted bandwidth. A well-behaved class offers traffic at a
rate no greater than the agreed-upon rate. By contrast, a misbehaving class offers traffic that exceeds the agreed-
upon rate. A misbehaving class is formed from an aggregation of misbehaving microflows. To achieve high link
utilization, a misbehaving class is allowed to use any idle bandwidth. However, such leniency must not degrade the
quality of service (QoS) received by well-behaved classes.
As Table 6 illustrates, the six traffic types may each have their own distinct properties and applications. As shown,
classes may receive bandwidth assurances or latency bounds. In the table, P3, the highest transmission class,
requires that all frames be transmitted within 1 ms, and receives 50% of the 100 Mbps of bandwidth at that port.
Best-effort (P0) traffic forms a fourth class that only receives bandwidth when none of the other classes have any
traffic to offer. It is also possible to add a fourth class that has strict priority over the other three; if this class has
even one frame to transmit, then it goes first. In the MVTX2602, each 10/100 Mbps port will support four total
classes and each 1000 Mbps port will support eight classes. We will discuss the various modes of scheduling these
classes in the next section.
In addition, each transmission class has two subclasses, high-drop and low-drop. Well-behaved users should rarely
lose packets. But poorly behaved users – users who send frames at too high a rate – will encounter frame loss and
the first to be discarded will be high-drop. Of course, if this is insufficient to resolve the congestion, eventually some
low-drop frames are dropped and then all frames in the worst case.
Table 6 shows that different types of applications may be placed in different boxes in the traffic table. For example,
casual web browsing fits into the category of high-loss, high-latency-tolerant traffic, whereas VoIP fits into the
category of low-loss, low-latency traffic.
Goals
Total Assured
Bandwidth
(user defined)
Low Drop Probability
(low-drop)
High Drop Probability
(high-drop)
Highest transmission
priority, P3
50 Mbps
Apps: phone calls, circuit
emulation.
Latency: < 1 ms.
Drop: No drop if P3 not
oversubscribed.
Apps: training video.
Latency: < 1 ms.
Drop: No drop if P3 not
oversubscribed; first P3 to drop
otherwise.
Middle transmission
priority, P2
37.5 Mbps
Apps: interactive apps, Web
business.
Latency: < 4-5 ms.
Drop: No drop if P2 not
oversubscribed.
Apps: non-critical interactive apps.
Latency: < 4-5 ms.
Drop: No drop if P2 not
oversubscribed; firstP2 to drop
otherwise.
Low transmission
priority, P1
12.5 Mbps
Apps: emails, file backups.
Latency: < 16 ms desired,
but not critical.
Drop: No drop if P1 not
oversubscribed.
Apps: casual web browsing.
Latency: < 16 ms desired, but not
critical.
Drop: No drop if P1 not
oversubscribed; first to drop
otherwise.
Total
100 Mbps
Table 5 - Two-dimensional World Traffic