13

To transmit data you need a media (copper or optical fiber, for example) and each media has its own set of features than make it usable for data transmission. For example copper is a good electrical transmitter and it's cheap (compared to other elements as silver or gold). Then, the best way to use it is to shape it as a cable and transmit electricity ...


12

Label itself is either an aggregate label, which means label does not have rewrite information attached to it, so it does not know egress interface nor egress MAC address. Aggregate labels are used for example to connected networks. Aggregate label implies that you do not know egress information after MPLS lookup, so you must do normal IP lookup to determine ...


9

Of course the NIC can be a bottleneck, if you have a NIC that can only handle a link of 100mbps, you can only transfer data at a max of 100mbps through that NIC - regardless of the other hardware specs of any hardware on the network. Same for if you have a 1Gbps NIC, the max throughput through that single NIC can only be 1Gbps. Your network is only as fast ...


7

Pushing full routes into a load balancer is about as far from best practice as you can get. Give it a default route to an actual router and a static route back to your subnets and it should be able to do anything you'd want. This also leaves you free to use that actual router to terminate multiple connections for diversity / best path options without ...


7

These are not different types of routers, they are different roles. It usually depends on where, topologically, the device placed in the network. Edge routers, for example, are placed at the network edge, while a core router is, well..., in the core. Routers all do the same thing: they forward packets based on layer 3 information. they might have ...


6

MTU is typically a reference to the physical medium of the directly connected interface. As a general rule, you want to leave this as the maximum the medium can handle. There are occasional cases where you need to enable jumbo frames or adjust MTU due to overlay networks or fancy encapsulation schemes. MSS is mostly a reference to the end-to-end path. ...


6

(Disclaimer: I am not familiar with HP switches, so I wont' be able to give proper command syntax for show/display commands). For a first, let's leave all routing discussion aside. AS5500 an AS3100 are on a common subnet (172.16.1.0/24) on VLAN1, so they must be able talk to each other directly. Accessing AS3100 from "elsewhere" is for later, we first have ...


6

Yes. You do not want to extend Layer 2 MAN links beyond the border of each location - this is bad design for a number of reasons: You will extend the broadcast domain of all 5 sites up to the main site Layer 2 issues (STP re-convergence, broadcast storms etc) can now affect multiple sites at once It will not scale in the future as you bring on more sites, ...


5

I am not able to find information on BHH; I got this link to draft-bhh-04 but was not able to understand the operation/function of BHH. We are stepping into territory that is very divisive and opinionated. My intent is just to provide information about the question. A very good overview of all the issues (from a US carrier's perspective) can be found in ...


5

The layer you labeled in your diagram is considered the core (also called the backbone). The firewalls in your diagram form a boundary on the core layer, since they transition to another network. Core networks are local to your company or autonomous system. Keep your core layer as stable as possible. That typically means keeping changes to a minimum, ...


5

To add to the previous comment, there are two MTU settings in the network itself that needs to be set. There is the Ethernet MTU , which you set on the switches. Again, set this to the maximum permissive value. Then there is the IP MTU you set on routers. Typically, if there are no options set (like the DF bit), then this just means the router will fragment ...


5

I think your question is not truly correct. First of all, as explain here "an MPLS VPN is a VPN that is built on top of an MPLS network, usually from a service provider, to deliver connectivity between enterprise office locations. " The MPLS is the mechanism that directs data from one network node to the next, based on short path labels instead network ...


5

There are quite a few global networks other than internet, mostly used for scientific research or military / diplomatic needs. They may or may not use IP as the needs of their users dictate. For example, you can find the current map of the Global Lambda Integrated Facility (GLIF) here, which is one such network. Internet2 is also a separate initiative, ...


5

Proceeding from the assumption that you are talking about Cisco gear, it is basically as you said, they complete each other. Neither an access-list nor an access-group on its own will have any effect. In fact, you configure an access-list first, which contains the rules permitting or denying the traffic. You can match on things like source and destination ...


5

Routing works the same in IPv6 as IPv4, so yes, a /128 address can be routed. However, you should not create subnets with prefixes longer than /64**. RFC 7421 explains why. ** The exception is point to point links using a /127, as described in RFC 6164


5

Just so we're clear on terminology, in the 802.11 spec, Ad-Hoc is a mode of operation of the wireless client interface - this allows direct connectivity between two (or more) wireless stations that have the same SSID configured. This is as opposed to Infrastructure mode in which the wireless interface or station must be able to associate to an Access Point ...


5

It's the last point of significance in a service provider's network, relative to the consumers of that service, from the point of view of the service provider. For example, in the days of dial up Internet service, the consumers had modems in their homes attached to phone lines, and the head end was the concentrator where all those phone lines met up. in ...


5

I think he's referring to "zero downtime" if you're connected to multiple switches. If you're only connected to Switch A, and I reload that switch for a software update, you're obviously going to lose connectivity. However, with multi-chassis-LACP, you have multiple connections to the "fabric", so a reload of a single switch will not cause a drop. This is ...


5

If you need to forward between different L2 segments/L3 subnets then any device actually doing that is a router by definition. So, there's no alternative to a router (which may be a classic router, software on a host, a layer-3 switch, a routing firewall, SDN hardware, ...).


4

Anything network traffic flows through can be a bottleneck. Large, popular websites usually use multiple servers behind load balancers to, among other things, keep the NIC(s) on one server from becoming a bottleneck. You need to look at every point in a network to see where the bottlenecks are. For instance, many people will install a 48-port switch with 1 ...


4

You should announce both networks to each eBGP peer and work with a route-map to prepend your AS path for influencing the incoming traffic on both routers. (of your EBGP have maybe a BGP community?)


4

It should be calculated based on the respective fields in the respective information storage of HSS/MME/SGW/PGW as mentioned in section 5.7. One tip is that the following 3GPP specs can be referred for respective IE in information storage to determine the sizes. S1AP IEs - 36.413 RRC IEs - 36.331 GTP-C IEs - 29.274 GTP-U IEs - 29.281 NAS IEs - 24.301


4

They have described the technology ("a really long Ethernet cable") pretty accurately. Typically transport in this case is referring to "lit services" by a provider. The technologies used will vary widely depending on a number of factors, including the provider, the geographic area, etc. but the most common instances I've seen are the following: 1) The ...


4

By contradiction: Suppose p is the shortest path across the graph and p has a loop. Remove the loop. The new loop-less path across the graph is shorter than p. We have achieved a contradiction, so our initial supposition must be false. (This applies only if "shortest" means "smallest sum of edge weights" and there are no negative edge weights.)


4

What is driving CPU usage in a Cisco switch is probably not a single end-device. Cisco switches can normally handle a full switching load without driving the CPU utilization unusually high. This is because the switching is largely handled in hardware. The most likely cause is a service or protocol which the switch needs to handle in software. You didn't ...


4

IP is a layer-3 protocol, and TCP is a layer-4 protocol. Neither protocol understands files. An application will send data to layer-4 (TCP), and TCP will break the data into chunks and encapsulate it in TCP segments. TCP will pass the segments to layer-3 (IP), which in turn encapsulates the segments in packets. IP will pass the packets to layer-2 (ethernet), ...


4

BGP does not care if the peer is directly attached or not, as long as peers can communicate with each other, so from BGP point-of-view both topologies are strictly identical. This is why there's an IGP running within the AS so each BGP router knows how to reach its peers. The fact that the IGP can be iBGP may cause confusion. Design your iBGP topology as ...


4

You are confusing MTU and MSS. The MTU 1500 is related to ethernet, so it includes the IP and TCP headers. Both have a minimum size of 20, so the MSS (maximum segment size) for TCP is in this case 1500 - 20 - 20 = 1460. Note that both the IP header and the tcp header can have a size greater than 20, when using options, and so you may still encounter ...


4

A flow is the data-plane stream of packets between sender and receiver that shares key IP header information. For example, a client at 10.1.1.1 port 12398 communicating with a server at 192.168.1.1 port 22 for SSH is a specific flow that can be captured as the key fields don't change. A session is the control-plane communication between sender and receiver....


4

You should check the portal server limitations, since it is enabled on A5500 Vlan 1 interface: # interface Vlan-interface1 ip address 172.16.1.1 255.255.255.0 portal server 1 method layer3 # While a portal server hasn't stateful packet inspection functionality, it could prevent icmp echo reply response to pass Vlan 1 layer 3 interface, when directed to ...


Only top voted, non community-wiki answers of a minimum length are eligible