- Posts: 9
- Thank you received: 0
STUDY GUIDE CCNP-ONT ( 642-845 )
15 years 3 days ago #32692
by beibei520
STUDY GUIDE CCNP-ONT ( 642-845 ) was created by beibei520
Priority Queuing
You can't earn your CCNP certification without passing the Cisco ONT exam, and queuing is a huge topic on that exam! Priority queuing, LLQ, weighted fair queuing - you'll see them all and more on your ONT exam. Today, we'll take a close look at priority queuing theory.
Priority queuing is unique in that it has four pre-configured queues, and while we have some control over those queues, we can't add more of them. Here are the four queues and their default capacity. Each capacity shown here can be changed.
High Priority queue, 20-packet capacity.
Medium Priority queue, 40-packet capacity.
Normal Priority queue, 60-packet capacity. (This is the default queue for all traffic when PQ is in use.)
Low queue, 80-packet capacity.
It's up to the network administrator to configure what types of traffic will be placed into each queue, and the key to success with PQ is not defining too many traffic types as high priority. It's vital to remember that PQ is *not* fair, and does *not* work in a round-robin fashion. When packets arrive in the High queue, PQ drops everything it's doing in order to transmit those packets, and packets in other queues are ignored until the High queue is again empty.
If you have too many packet types being placed into the High queue, and even the Medium queue, traffic in the lower-priority Normal and Low queues ends up just sitting there. That's called queue starvation or packet starvation, but whatever you call it, it's a danger with priority queuing - a danger you must avoid!
That's priority queuing theory; now we need to work on some configurations, and we'll do just that in the next installment of this CCNP certification training series!
QoS Models
A large part of your CCNP training for the ONT certification exam should be spent studying the various ways we can implement Quality of Service (QoS) on Cisco routers and switches. Before you start configuring your network's devices, though, you've got to understand the three QoS models and their impact on your network. Let's take a look at the pros and cons of each.
If you don't have a QoS model in place, you actually do. Best-effort QoS is just that - best-effort. No priority is given to any traffic. If your network is carrying voice or video traffic, best-effort is definitely not the way to go.
The Integrated Services model, more popularly known as IntServ, uses the Resource Reservation Protocol (RSVP) to reserve network resources in advance of the data actually traveling across the network. Once the end-to-end bandwidth reservation is in place, the data is transmitted.
That sounds great, but there are some drawbacks. It's a waste of bandwidth to have the entire end-to-end path reserved in advance. Additionally, IntServ isn't as scalable a solution as we'd like. Everything we do on a router or switch has a cost of some kind, and in this case it's RSVP overhead. One or two paths won't cause much overhead, but as the number of reserved paths increases as a network becomes larger, the RSVP overhead can take its toll on the routers involved.
Differentiated Services (DiffServ) is the latest of the three models, and many would agree that it's also the greatest. DiffServ doesn't use RSVP, but instead uses Per-Hop Behavior (PHB) to allow each router across the network to examine the packet and decide what service level it should receive. With DiffServ, one router along the path from source to destination could consider a packet to be of the highest priority, while another router could consider it "just another packet". A term you hear often with DiffServ is "marking and classification". Marking a packet is the process of assigning the packet a value reflecting the level of QoS it should receive, while classification is placing that packet into a queue in accordance with that level of QoS.
When it comes to marking, there are different values we can use to decide what value to mark the frame or packet with. In my experience, here are the four that are used most often:
* IP Precedence (IP Prec) * Differentiated Services Code Point (DSCP) * CoS value * Interface that received the data (ingress interface)
Which one you choose depends on your particular network's needs, and of course, the OSI layer at which the marking is taking place. We'll take a look at each of these methods in future CCNP ONT exam training tutorials!
more ccnp informationwelcome to passcert
You can't earn your CCNP certification without passing the Cisco ONT exam, and queuing is a huge topic on that exam! Priority queuing, LLQ, weighted fair queuing - you'll see them all and more on your ONT exam. Today, we'll take a close look at priority queuing theory.
Priority queuing is unique in that it has four pre-configured queues, and while we have some control over those queues, we can't add more of them. Here are the four queues and their default capacity. Each capacity shown here can be changed.
High Priority queue, 20-packet capacity.
Medium Priority queue, 40-packet capacity.
Normal Priority queue, 60-packet capacity. (This is the default queue for all traffic when PQ is in use.)
Low queue, 80-packet capacity.
It's up to the network administrator to configure what types of traffic will be placed into each queue, and the key to success with PQ is not defining too many traffic types as high priority. It's vital to remember that PQ is *not* fair, and does *not* work in a round-robin fashion. When packets arrive in the High queue, PQ drops everything it's doing in order to transmit those packets, and packets in other queues are ignored until the High queue is again empty.
If you have too many packet types being placed into the High queue, and even the Medium queue, traffic in the lower-priority Normal and Low queues ends up just sitting there. That's called queue starvation or packet starvation, but whatever you call it, it's a danger with priority queuing - a danger you must avoid!
That's priority queuing theory; now we need to work on some configurations, and we'll do just that in the next installment of this CCNP certification training series!
QoS Models
A large part of your CCNP training for the ONT certification exam should be spent studying the various ways we can implement Quality of Service (QoS) on Cisco routers and switches. Before you start configuring your network's devices, though, you've got to understand the three QoS models and their impact on your network. Let's take a look at the pros and cons of each.
If you don't have a QoS model in place, you actually do. Best-effort QoS is just that - best-effort. No priority is given to any traffic. If your network is carrying voice or video traffic, best-effort is definitely not the way to go.
The Integrated Services model, more popularly known as IntServ, uses the Resource Reservation Protocol (RSVP) to reserve network resources in advance of the data actually traveling across the network. Once the end-to-end bandwidth reservation is in place, the data is transmitted.
That sounds great, but there are some drawbacks. It's a waste of bandwidth to have the entire end-to-end path reserved in advance. Additionally, IntServ isn't as scalable a solution as we'd like. Everything we do on a router or switch has a cost of some kind, and in this case it's RSVP overhead. One or two paths won't cause much overhead, but as the number of reserved paths increases as a network becomes larger, the RSVP overhead can take its toll on the routers involved.
Differentiated Services (DiffServ) is the latest of the three models, and many would agree that it's also the greatest. DiffServ doesn't use RSVP, but instead uses Per-Hop Behavior (PHB) to allow each router across the network to examine the packet and decide what service level it should receive. With DiffServ, one router along the path from source to destination could consider a packet to be of the highest priority, while another router could consider it "just another packet". A term you hear often with DiffServ is "marking and classification". Marking a packet is the process of assigning the packet a value reflecting the level of QoS it should receive, while classification is placing that packet into a queue in accordance with that level of QoS.
When it comes to marking, there are different values we can use to decide what value to mark the frame or packet with. In my experience, here are the four that are used most often:
* IP Precedence (IP Prec) * Differentiated Services Code Point (DSCP) * CoS value * Interface that received the data (ingress interface)
Which one you choose depends on your particular network's needs, and of course, the OSI layer at which the marking is taking place. We'll take a look at each of these methods in future CCNP ONT exam training tutorials!
more ccnp informationwelcome to passcert
Time to create page: 0.109 seconds