Can OpenWrt Route



Nor is ping permitted to this device from another:

root@OpenWrt:~# ping
PING ( 56 data bytes
ping: sendto: No error information

Does it not work because OpenWRT won't initialize the devices correctly (due to restrictions in the config file stuff) or because the devices are initialized correctly but the kernel prevents it from working?

If you manually set up the device with "ip" does it work?

1 Like

Hi, lleachii:

  1. "Each IP uses about 65,000 TCP and UDP ports, if you alter either the IP/port of the SRC or DST, you have broken the end-to-end principle.":

If you are referring to this, my answer is No. SPR does not alter TCP / UDP port numbers as shown by the examples in the EzIP Draft Appendices.

  1. "End-to-end principle":

I came from the old telephony discipline where this principle was even more carefully observed than in the Internet, because the analog signals were more prone to degradation. So, I believe that EzIP has seriously carried such disciplines on.

  1. "But, I assume this "space" will be controlled by ISP routers, they need to be able to route as well. ":

No for sure. The RAN as I described looks like a private network to the ISP (Internet), because it appears like a few IoTs tethering off a mobile phone on one IPv4 address. I hope you agree that the ISP has not much to say about those IoTs in such configurations. Then, when the number of these IoTs becomes huge (256M), the ISPs still should not have anything to do about how they are interconnected. But, this RAN can physically cover a lot of area overlapping the existing territories of the ISP. Essentially, the current Internet will be "marginalized" to only serve traffic between RANs, This is very much analogous to how electric utility grid providing the backup to islands of renewable energy generated by businesses and homes.

Hi, Colleagues:

Great! I am glad that you are getting into the actual test with your own equipment.

It seems that dlakelan says that it works while lleachii is getting no.

Allow me staying on the sideline to monitor your interactions. When your reach a definitive approach, I should dive into trying your recommendation.

Thanks a lot.

1 Like

it works at the kernel level. I added to my ethernet device and it tries to ping (but finds nothing, because there is no other device on that network to respond). I can ping since it's local. This is all on a desktop linux machine running Debian and kernel 4.16.0 and adding the addresses as "ip addr add dev eno1" and "ping" (arp doesn't work because this device doesn't exist) or "ping" (works fine, since it's a local address) i'm using "iputils-ping" for pinging.

EDIT: Behavior is the same when I log into an OpenWRT router and manually add a address using ip. I am not willing to reconfigure my in-use access points etc. and I don't have a test device to test it on, but if @lleachii is seeing some other behavior it's probably due to some UCI config checking and refusing that address range rather than a fundamental problem in the kernel or other binaries.

1 Like

Why don't you spend $20 and buy a device yourself?

Long ago you were told that OpenWrt ran a Linux kernel, along with how to modify that kernel and standard Linux-based OS networking utilities.

If you (signed as "VP of Engineering") and your "seasoned IT professionals" can't configure a basic Linux-based OS to test, then I am going to consider this all link and keyword spam in an attempt to promote your commercial efforts.


That your firm's website basically only includes 2001, 2002 press releases, your firm's "about" page provides a Sunnyvale location (not Milpitas), and you have no viable physical products listed, and that the technology products listed link to broken pages (404 on the page itself, as well as within your document handler) on your firm's website adds to the belief that that this is little more than marketing or SEO attempts, attempting to leverage the high reputation of the OpenWrt domain.


Funny, Manta lists Abraham T Chen as the Chief Executive Officer of your firm, not the "VP of Engineering". Your own words, "my two coauthors" on the draft pretty clearly identify that name with you.


Even more interesting, in the metadata of your company's products page is "[your company] is a systems engineering company specializing in voice and data networking technologies and system architectures. [your company] licenses intellectual properties for products that enhance consumer's networking experiences"

So, you don't really build anything, do you?


If the end-to-end principle is never broken, I misunderstand the need for your EzIP technology.

   a new category named Semi-Public Router (SPR). By inserting an SPR
   between an ER and a private premises that it serves, each publicly
   assignable address is expanded by 256M fold.
  • How do you do so, without breaking the end-to-end principal?

You examples never show how:

  • the packet returns to the unmodified host from a downstream EzIP enabled server

Assume - I'm a customer in your EzIP "cyberspace" - trying to reach standard Global IPs:

  • How do I run a standard webserver on port 80 when there are 256M others running webservers on the same Public IP "NATed" to IP at TCP 80 (EXPLAIN ONLY USING LAYER 3 OF THE OSI MODEL)!?!?
  • In the does a client running a protocol requiring a TCP or UDP packet be sent with SRC port x - enter and exit your SPR to the Global Internet if that TCP port is in use? (assume all clients are trying to reach the same Global IP using the same protocol/port combination)

by manually making the selection from the main web server.

  • This is at least a Layer 7 function, why does this protocol break the OSI and DARPA network models???
  • This is solved in large web operations all the time, with Private IP networking and proxy servers

@OugCPC, all the best in your endeavor!

1 Like

As if there wasn't enough already, the address given for the OP's "company" is a UPS Store in a shopping center.

Yes, the storefronts are individually numbered, check "Nail Nook, Milpitas", for example.

@jeff, I wondered how they came to be in the same place as:


I must indulge...I missed that...I really think you need to read the RFCs, starting with:

eventually you wil get to:

You will find:

4. HOST GROUP ADDRESSES Host groups are identified by class D IP addresses, i.e., those with "1110" as their high-order four bits. Class E IP addresses, i.e., those with "1111" as their high-order four bits, are reserved for future addressing modes. In Internet standard "dotted decimal" notation, host group addresses range from to The address is guaranteed not to be assigned to any group, and is assigned to the permanent group of all IP hosts (including gateways). This is used to address all multicast hosts on the directly connected network. There is no multicast address (or any other IP address) for all hosts on the total Internet. The addresses of other well-known, permanent groups are to be published in "Assigned Numbers".

You will then eventually find: which tells you how to find the current publication of "Assigned Numbers."

You then find yourself BACK HERE:

I was very serious when I said:

:thinking: Do you wonder why you haven't been issued a number???

Understatement. You have something that is blatantly false in your draft!

   7. IANA Considerations

   This draft does not create a new registry nor does it register any
   values in existing registries; no IANA action is required.
1 Like

@OugCPC are aware that the IETF mostly controls the RFCs...with the ISOC...which I'm a member...???:

The IETF is overseen by the Internet Architecture Board (IAB), which oversees its external relationships, and relations with the RFC Editor.[6] The IAB is also jointly responsible for the IETF Administrative Oversight Committee (IAOC), which oversees the IETF Administrative Support Activity(IASA), which provides logistical, etc. support for the IETF. The IAB also manages the Internet Research Task Force (IRTF), with which the IETF has a number of cross-group relations.


EDIT: All peoples can become a member of the ISOC

Hi, Jeff:

  1. " Why don't you spend $20 and buy a device yourself? ":

Now we are talking basics. My dumb question should have been

A. "What lowest cost retail router(s) that you could recommend that would route the netblock?

B. To make the result available to the largest possible group of dummies, could you correlate your recommendation to the offering from two lowest cost CE outlets that we know of?

  1. "your firm... ":

This is getting interesting. I am learning the culture of this forum, because certain member seems to push anonymity. Yet, you are now digging into our company's details. To clear the subject once for all so that we can focus on the task on hand, Avinta was a SOHO product manufacturer. You should find a product called VN100 / DP100 with sufficient information to tell you what it is. (By the way, we stopped making it quite sometime ago because the price/cost ratio. But, we continue to get calls for it. Some customers even came back to ask for more. .... ) We then evolved into an international R&D organization, maintaining the registration in California, USA. What you found is our mailing address. Our main efforts in recent years led to the EzIP proposal. The other information that you read were accumulated there for various past situations. It will be wasteful for me to go through the details.

Shame...I really want to know why:

  • A company in Milpitas
  • Which is listed on a draft RFC
  • Proposes an IPv4 exhaustion technology other than CGN or;
  • IPv4's replacement, IPv6
  • Deflects all comments as: "marketing" IPv6, etc...
  • Why this person seem to not understand the end-to-end principle
  • And then gets quite despondent
  • And doesn't seem to understand he filed an RFC, and just received "Comments"


Hi, lleachii:

  1. "How do you do so, without breaking the end-to-end principal? ":

The SPR is a plain inline router that provides clear channel for either direction of a session setup request.

  1. " You examples never show how: the packet returns to the unmodified host from a downstream EzIP enabled server ":

It is explained in Appendix A. 3. that the initiating IoT determines which mode of operation that SPR will provide. If it does not send out a packet with EzIP header, everything (including those EzIP-capable) in the loop uses IP header.

In Figure 7, SPR1 allowing masqueraded by RG1


You are breaking the end-to-end principle!

Hi, lleachii:

  1. "Condemned!?!? ":

This word is used as figure of speech, meaning the 240/4 block is not usable anywhere in the normal Internet operation. This is similar as condemning a house by declaring it unfit for occupancy.

  1. " I was very serious when I said: ":

Thanks for your comments. It is getting too much intertwined politics which I should refrain from making any comments on your short descriptions.

  1. The key issue I believe is whether you can see that a RAN may be stealthily deployed independent of all these "regulatory" bodies? This is a system architectural question that must precede any of the technical considerations. If not, I am not surprised that you have so many questions. missed a word there that tells me to shut up. We'll see if the IANA says OK. As you know...RFC1112 reserves that space. If/when it ever changes...

Hope this helps....

(...still breaks the end-to-end principle...caveat emptor)


Closing due to this topic increasingly getting off-topic. Answer to the original question is yes, OpenWrt, like Linux, can route arbitrary CIDRs when configured to do so.