Review of the new line of Cisco Fatalyst switches
Literally, a new series of switches from one of the largest vendors, Cisco Systems, has just been announced.
In addition to the Nexus and Catalyst lines, a new Cisco Fatalyst line will be released, and, accordingly, the approach to building local networks has been revised. The concept of Borderless Networks is being replaced by a new concept of Hopeless Networks , the main support in it is exactly on L2, L3 and lvl80 Fatalyst switches.
')
Especially for those who have tl; dr, at the end of the article describes the conditions for obtaining two switches for temporary testing.
So, what is interesting in the new switches from Cisco:
A little more about routing. A number of improvements have been announced, which made it possible to achieve performance of more than 9000 Gbit / s while routing to Null0.
By default, routing-enabled switches do this based on destination (so-called destiny-based routing ). It is also possible to use PBR ( prophecy-based routing ).
Improvements also affected routing protocols. Those who studied EIGRP remember that one of the “chips” of the protocol is the DUAL algorithm and the operation of a finite state machine (Finite State Machine). Cisco engineers went further and improved the algorithms by using the so-called. Only-State Predeterministic Final State Machine (of a finite predefined automaton with a single state). Scheme of operation of the machine is shown in the figure below:
This machine allows you to get stable routes without cycles for loopback interfaces. For backward compatibility, the NOHOPE (No Overhead over Plain EIGRP) standard is used.
The changes in IPv6 primarily affected the SLAAC algorithm (also known as the Stateless address configuration). This approach has a number of weaknesses, for example, confusion and potential problems due to constantly changing temporary addresses. On Fatalyst switches, the SLAAC mechanism has been replaced by HLAAC ( Hopeless autoconfiguration ) with SAD ( Spontaneous address detection ) support. A new way to get addresses ensures that regardless of the situation and settings, the switch will always receive the same predefined address. In addition, at the expense of SAD, the assignment of any other (“spontaneous”) addresses to any interfaces is excluded.
Other improvements include support for PPP ( Pointless-to-pointless protocol ) and the ability to enable depression instead of compression on interfaces.
And finally, the most significant change lies in the way the device is configured.
First, the switches will come to customers already configured. This decision is intended to improve the situation with the deployment of systems, especially in the conditions of a shortage of qualified personnel. Now the device will be configured “out of the box”, and there’s nothing to do at all.
Secondly, Cisco's research showed that a very large proportion of failures and downtime is not at all on the conscience of the failed equipment, but on the conscience of network engineers who change the correct settings to erroneous, that is, human factor (did you set up logging of entered commands?). For this reason, the configuration on Cisco Fatalyst cannot be changed. No It is predetermined.
Special mention deserves the fact that Cisco has promised in the near future to publish hopeless software.
If you are interested in this, you need to write in a personal message-essay, in which to highlight two topics:
1) How Cisco Fatalyst Switches and Hopeless Networks Concept Will Not Change the World and
2) Why do you want to get them for testing (note that, for example, "beautiful" and "yes, for fun" are not valid reasons).
We reserve the right to subsequently publish the essays we like most at the end of this article.
In preparing the used information from the datasheet ( link )
For the idea of ​​post thanks evilenka!
In addition to the Nexus and Catalyst lines, a new Cisco Fatalyst line will be released, and, accordingly, the approach to building local networks has been revised. The concept of Borderless Networks is being replaced by a new concept of Hopeless Networks , the main support in it is exactly on L2, L3 and lvl80 Fatalyst switches.
')
Especially for those who have tl; dr, at the end of the article describes the conditions for obtaining two switches for temporary testing.
So, what is interesting in the new switches from Cisco:
- The complete set of all models is fixed, it cannot be changed in any way (even SFP). In the future, a line of modular switches will probably appear, but even there the modules are most likely to be rigidly fixed in slots.
- New design - black color ( gravely black ) in combination with the case of a special alloy (" doom metal "). The model Fatalyst EMO waterproof case, with inserts in pink.
- Significant improvements in routing and IPv6 support (see below).
- Great for networks in the context of Internet of Things with Sustainable Cloud SDN of Everything .
A little more about routing. A number of improvements have been announced, which made it possible to achieve performance of more than 9000 Gbit / s while routing to Null0.
By default, routing-enabled switches do this based on destination (so-called destiny-based routing ). It is also possible to use PBR ( prophecy-based routing ).
Improvements also affected routing protocols. Those who studied EIGRP remember that one of the “chips” of the protocol is the DUAL algorithm and the operation of a finite state machine (Finite State Machine). Cisco engineers went further and improved the algorithms by using the so-called. Only-State Predeterministic Final State Machine (of a finite predefined automaton with a single state). Scheme of operation of the machine is shown in the figure below:
This machine allows you to get stable routes without cycles for loopback interfaces. For backward compatibility, the NOHOPE (No Overhead over Plain EIGRP) standard is used.
The changes in IPv6 primarily affected the SLAAC algorithm (also known as the Stateless address configuration). This approach has a number of weaknesses, for example, confusion and potential problems due to constantly changing temporary addresses. On Fatalyst switches, the SLAAC mechanism has been replaced by HLAAC ( Hopeless autoconfiguration ) with SAD ( Spontaneous address detection ) support. A new way to get addresses ensures that regardless of the situation and settings, the switch will always receive the same predefined address. In addition, at the expense of SAD, the assignment of any other (“spontaneous”) addresses to any interfaces is excluded.
Other improvements include support for PPP ( Pointless-to-pointless protocol ) and the ability to enable depression instead of compression on interfaces.
And finally, the most significant change lies in the way the device is configured.
First, the switches will come to customers already configured. This decision is intended to improve the situation with the deployment of systems, especially in the conditions of a shortage of qualified personnel. Now the device will be configured “out of the box”, and there’s nothing to do at all.
Secondly, Cisco's research showed that a very large proportion of failures and downtime is not at all on the conscience of the failed equipment, but on the conscience of network engineers who change the correct settings to erroneous, that is, human factor (did you set up logging of entered commands?). For this reason, the configuration on Cisco Fatalyst cannot be changed. No It is predetermined.
Special mention deserves the fact that Cisco has promised in the near future to publish hopeless software.
How to get a couple of Cisco Fatalyst switches for testing
If you are interested in this, you need to write in a personal message-essay, in which to highlight two topics:
1) How Cisco Fatalyst Switches and Hopeless Networks Concept Will Not Change the World and
2) Why do you want to get them for testing (note that, for example, "beautiful" and "yes, for fun" are not valid reasons).
We reserve the right to subsequently publish the essays we like most at the end of this article.
In preparing the used information from the datasheet ( link )
For the idea of ​​post thanks evilenka!
PS
Happy April, 1!
Source: https://habr.com/ru/post/217787/
All Articles