30

CAM (Content Addressable Memory) is memory that can be addressed by content, rather than a numeric memory address. You can look up the interface by presenting the memory with the MAC address. This is done in a single CPU cycle vs. the traditional programming of searching through a table, which will cost many CPU cycles. There is also TCAM (Ternary Content ...


20

No performance gain will exist without everyone using larger packets. The point of jumbo frames is to pack more payload with the same overhead. The NAS's ability to send larger packets is meaningless if the clients don't as well. There will be no "fragmentation" at all. Layer-2 (ethernet) has no means if indicating "fragmentation needed". This is figured ...


20

Is it correct in assuming that ASICs for router or switch interfaces will outperform the use of an x86 CPU for all packet processing which will greatly suffer from CPU interrupts? It's hard to say specifically whether interrupts are a limitation, since we aren't naming specific CPU, operating system or router models in this part of your question. Overall, ...


18

Yes. A managed switch is a switch you can configure in some way or other. Whether it supports VLANs or not is not the question. Even a switch (or a hub for that matter) that only provides status information can be considered "managed". However, the vast majority of managed Ethernet switches do support VLANs. There are countless other features than can be ...


14

CAM - Content Addressable Memory, referring to the memory used for the MAC address table. It works kind of reverse from RAM, you address it by giving it content and it returns you the address where the content is stored - which is then used to find the egress port for this address.


14

Good question. The short answer: No, there's no inherent difference in the speed or latency available to hosts talking to one another on a managed vs. unmanaged switch. In reality though, you'll generally see better performance all around on a managed switch, because unmanaged switches are typically less capable in every sense when compared to a managed ...


13

Generally, a giant frame is a frame that is too large for the receiving interface. As a malformed frame it is dropped. A jumbo frame is a frame that is larger than the standard allows (1518 bytes for Ethernet w/o tags, or 1500 bytes L3 payload (= L3 PDU = L2 SDU) plus L2 overhead). It may still be acceptable, depending on the interface configuration. For ...


12

You can certainly use Port Mirroring with the HP 2915. You can find the information for port mirroring on page B-23 in the documentation for the 2915. You will be able to designate monitoring of inbound and outbound traffic on either individual ports, groups of ports, static port trunks, and one static VLAN. An example to assign a port as the monitoring ...


12

No. If all the switches are layer-2 switches, the frames are switched without any changes. Only with routers, including layer-3 switches where the packets need to cross to other VLANs, will the frames be stripped and rewritten for the new network or VLAN.


12

You really, really do not want to disable STP where you connect switches to other switches. That is the entire purpose of STP. If you disable STP, and there is a problem, it will really be too late because your entire network could crash when you notice it, and recovering from a broadcast storm is no fun at all. By the way, portfast doesn't actually disable ...


12

You need to understand the concept of layers. An application will send data to the Transport Layer. The Transport Layer protocol will encapsulate the data inside headers for the Transport Layer protocol, and pass those to the Network Layer. The Network Layer will encapsulate the datagrams it receives inside Network Layer headers, and those are called ...


12

A switch learns the source MAC from the sender. If the destination is not in the CAM table, the switch floods the frame out all ports. So if the receiver never responds, the switch will never learn the receiver's MAC and it will always flood the frame.


11

The forwarding information base (FIB) is the actual information that a routing/switching device uses to choose the interface that a given packet will use for egress. For example, the FIB might be programmed such that a packet bound to a destination in 192.168.1.0/24 should be sent out of physical port ethernet1/2. There may actually be multiple FIB's on a ...


10

Most often Cisco devices can only receive PAUSE frames. They can't send them. If you are running storage over your network I can understand why you would be looking at implementing it and some server/storage vendors even recommend you to do so. Note however that PAUSE frames is a very blunt tool as it can pause all traffic meaning you can't differentiate ...


10

Port isolation -also called private VLAN (thanks @Stuggi)- is a very useful feature for switches that connect end users. In a typical network you will have many end-users computers grouped together in a VLAN that communicates with some servers in other networks. Those computers have no need to communicate together, so it's best to block those unwanted ...


10

It depends on the manufacturer. For Cisco switches: Switch port configured as access and frames received tagged on different VLAN Frame is dropped Switch port configured as trunk and frames received tagged on non authorized trunk VLAN Frame is dropped Switch port configured as trunk without native VLAN and frames received not tagged Frame ...


9

FPC's are similar to Linecards. When dealing with Juniper you will hear FPC more often. FPC's (Flexible PIC Concentrator) houses multiple PICs (Physical Interface Cards) which connect to the physical medium. Linecards are similar in that both FPCs and Linecards are inserted into a chassis device. In Cisco, Linecards are just that. It's a card that's slotted ...


9

One key use for per port MAC addresses on switches is for Spanning-tree BPDU's. These are Layer-2 multicasts with source MAC address of the egress switch port. I would have to brush up on other Layer-2 protocols such as TRILL and SPB, but they might also take advantage of a per port MAC. Does that help?


9

Transport of both UDP and TCP packets from one router to the next is done at the IP layer and solely based on the information at this layer. This means, that there is no distinction between UDP and TCP and in both cases change of path, congestion or router hickups can cause loss, duplication or reordering of packets. But, contrary to UDP, TCP can deal with ...


9

In general, forwarding refers to a device sending a datagram to the next device in the path to the destination, switching refers to moving a datagram from one interface to another within a device, and routing refers to the process a layer-3 device uses to decide on what to do with a layer-3 packet. A host sending data to another host through an ethernet ...


9

Devices in different subnets can communicate. That is the purpose of a router. Routers route packets between different networks. Even if devices in different networks are on the same layer-2 broadcast domain, you need a router to let the devices communicate at layer-3. That is because each host will compare the destination layer-3 address and its own layer-...


9

For the most part, no. Hubs are also known as repeaters - meaning that they basically echo whatever is coming in to a given port out all other ports in the domain. The idea is (or, more properly, was) to replicate the behavior of a shared piece of coaxial cable, meaning that the individual hosts had to assure that only one member of the LAN was ...


8

In this case, you are wrong. Since the link between the two switches is an access port, there is no VLAN tagging involved. Think of it this way, switch two is a unconfigured switch (i.e. operating like a dumb switch) so all ports are in VLAN 1. This would still provide connectivity as well.


8

The SSG-20 only runs ScreenOS, not JunOS. Juniper switches run JunOS, so this won't help you test your scripts and test configs. Personally, I'd look for a cheap 2nd hand low end model Juniper, but it all depends a bit on your budget.


8

Smallest switch would be an EX-2200c. I bought one for my home lab. It has 12 ports and most of the important features. You can even use virtual chassis functionality (with up to 2 or 4 devices). Junosphere might also be worth a look. And I guess you already know about the Junos Fasttrack program where you can get the training materials for the "lower" ...


8

* UPDATED Dec 2017 * Juniper renamed this product to the vSRX (virtual SRX). You can still download a trial copy - https://www.juniper.net/uk/en/products-services/security/srx-series/vsrx/ You can also download Juniper Firefly which is a virtual SRX device (not the same as an Olive). It has a trial 60 day license so you can test out configuring ports and ...


8

It usually comes down to options and resources, where routers have more of each. Routers often have the ability to have different types of interfaces that are not available on switches. They also typically have more resources (RAM, processor, etc.). Routers also usually have a hardware assist for NAT, which is actually resource intensive, and most switches ...


8

An unmanaged switch will only have one VLAN. Some unmanaged switches will drop tagged frames as damaged, others will strip the tag, and some will simply pass the frames unchanged. Unless you try it, you have no way to know how your switch acts. For the switches that strip the tags, all the frames will be placed into the native (untagged) VLAN as they pass ...


7

802.1q is the technical standard for VLAN tagging (http://en.wikipedia.org/wiki/IEEE_802.1Q). This standard includes the placement of a "VLAN Tag" inside the header of Ethernet frames. This tag allows a link to carry multiple VLANs, as long as both devices recognize 802.1q tagging, because the tag contains the VLAN ID that the traffic belongs to. Such a ...


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