256kbps is guarantee bandwidth for class SILVER while the interface is under congestion.
Under normal load, class SILVER can potentially use more than 256kbps of bandwidth.
"Is the 256 being set aside until use?"
No. (It's a guarantee that the SILVER class of traffic should be able to obtain at least that much bandwidth. [As Edison also notes.] Class bandwidth not used, is available to other traffic classes.)
One more question. The "available bandwidth" string under show int s0/0/0.
Is that the bandwidth currently available?
Hardware is GT96K with integrated T1 CSU/DSU
MTU 1500 bytes, BW 384 Kbit, DLY 20000 usec,
reliability 255/255, txload 13/255, rxload 248/255
Encapsulation FRAME-RELAY IETF, loopback not set
Keepalive set (10 sec)
LMI enq sent 4585, LMI stat recvd 4585, LMI upd recvd 0, DTE LMI up
LMI enq recvd 0, LMI stat sent 0, LMI upd sent 0
LMI DLCI 1023 LMI type is CISCO frame relay DTE
FR SVC disabled, LAPF state down
Broadcast queue 0/64, broadcasts sent/dropped 764/0, interface broadcasts 0
Last input 00:00:00, output 00:00:00, output hang never
Last clearing of "show interface" counters 12:44:07
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: Class-based queueing
Output queue: 0/1000/64/0 (size/max total/threshold/drops)
Conversations 0/46/128 (active/max active/max total)
Reserved Conversations 2/2 (allocated/max allocated)
****Available Bandwidth 24 kilobits/sec****
5 minute input rate 374000 bits/sec, 43 packets/sec
5 minute output rate 21000 bits/sec, 29 packets/sec
2052446 packets input, 2008412730 bytes, 0 no buffer
That value reflects the bandwidth available for reservation. It does not reflect the available bandwidth for flows traversing this interface.
Avaliable bandwidth means as the name implies. You are allowed to use 75% of bandwidth by default. 25% is allocated for class-default. Let's say.
- Link bandwidth = 1024 kbps
- You are allowed to use 768 kbps. 75% of 1024
- You have assigned 256 kpbs for SILVER class
As a result. Available bandwidth is 768 - 256 = 512
Am I clear?
To clarify what Toshi is saying, it's not really "allowed to use" but more so "allowed to define" with bandwidth statements. Actual traffic bandwidth utilization, for non-LLQ classes, isn't normally limited by the bandwidth statement.
On many router platforms, by default, you can not define bandwidth statements that sum more than 75%. (NB: default can be overridden, and limitations differ on diffent platforms and/or very recent IOSs which support HQF.)
Thanks for fixing my words. That's right.
Like you mentioned, Available bandwidth is the rest of bandwidth we can define. All depend on the bandwidth command you configured within the interface level.
5P! for you
Nice post, buddy.
Queick question regarding the default.
Can you elaborate a bit more on that default of 25%. Where is that standard written so I could read up on it?
rated your post, too
"You can configure class policies for as many classes as are defined on the router, up to the maximum of 64. However, the total amount of bandwidth allocated for all classes included in a policy map must not exceed 75 percent of the available bandwidth on the interface. The other 25 percent is used for control and routing traffic. (To override the 75 percent limitation, use the max-reserved bandwidth command.) If not all of the bandwidth is allocated, the remaining bandwidth is proportionally allocated among the classes, based on their configured bandwidth. "
The 25 % of the bandwidth is reserved for the important/critical traffic like L2 keepalives and other system critical frames/packets/routing traffics
The sum of all bandwidth allocation on an interface should not exceed 75 percent of the available bandwidth on an interface. The remaining 25 percent of bandwidth is used for overhead, including Layer 2 overhead, control traffic, and best-effort traffic.
If you need to allocate more than 75 percent for RSVP, CBWFQ, LLQ, IP RTP Priority, Frame Relay IP RTP Priority, and Frame Relay PIPQ, you can use the max-reserved-bandwidth command. The percent argument specifies the maximum percentage of the total interface bandwidth that can be used.
Please check out this link.
You can use a "max-reserved-bandwidth" command to override the default.
5P! for you Victor
I am interested with the remaining 25% as well. Its said that the remaining 25% is for control traffic also. My questions are:
1. By default, if we do not specify a particular class for control traffic, then it will be included in this 25% class-default right?
2. Regarding my first question, does this guarantee that the control traffic will be able to pass through a link when congestion occur? coz in my opinion, if we do it like this (doesn't specify a particular class for control), there's a possibility that the control traffic gets dropped or doesnt get through when the link is congested. Am I right?
3. On the other hand, if I specify a particular class for control traffic, so its okay then if i override the default bandwidth to be used to 100% by using the command "max-reserved-bandwidth 100", right?
I don't think that the 25% is used for the class-default. I think that class-default uses the left-overs of other classes if no bandwidth is specified.
I thought by default
75% of the bandwidth can be allocated
to LLQ/CBWFQ (including the default class)
I think that control traffic is automatically prioritorized by the software and is part of the 25%
can't edit my post anymore
I was wrong
The remaining 25 percent is used for other overhead, including Layer 2 overhead, routing traffic, and best-effort traffic. Bandwidth for the CBWFQ class-default class, for instance, is taken from the remaining 25 percent.
class-default traffic should go with the 75% of BW.
I may be wrong but there is no sense to have best effort traffic to compete with routing protocol packets.
So I agree with your first post.
To be added:
not all cisco platforms have this concept of a system queue for control traffic.
For example on C7500, C7200 and GSR 12000 we had to explicitly provide a class to protect routing protocol messages like BGP or OSPF.
the 25% is a rule of thumb that can be reduced on high speed links to a lower value freeing BW to be left for user traffic.
Hope to help
Bandwidth reservation depends on the platform and IOS version. On many Cisco routers, CBWFQ won't allow you to explictly define more than 75% of the bandwidth unless the default reserved bandwidth reservation is changed on the interface.
The reason for this is to reserve bandwidth for other traffic, including, but not limited to, class-default (when bandwidth isn't defined for it).
Since only LLQ's bandwidth statement limits bandwidth, any class can use up to 100% of available bandwidth (this also assumes there isn't a defined shaper or policer within the class).
On many platforms CBWFQ, control traffic, unless explictly matched in a defined class, will fall within class-default. Assuming max reserved bandwidth hasn't been changed, and also assuming WFQ is active with class-default, control traffic is often not adversely impacted by other traffic.