Default Network Address

Dear Colleagues:

  1. Can anyone please tell me where (e.g., what module in the source code) the default network address, is defined?

  2. Can this be changed?

  3. If so, is there any restriction as to which netblocks are usable and which ones are not?

Thanks and Happy Holidays!

Abe (2018-12-20 15:14)

fgrep -rI path/to/openwrt will show you all occurrences of that string within the source code.

If you're still looking at profiting from the vast, untapped, restricted IPv4 address space, a similar approach would be called for. still indicates that the netblock is Reserved.

egrep -rI '192\.168\.1\.1[^0-9]' will reduce some of the "noise" with addresses in the range.

Hi, jeff:

  1. Appreciate your instructions.

  2. I downloaded a snapshot copy of the source code openwrt-b53111a.tar.gz and unzipped it. When I attempted to apply your command lines from the same directory, the first one gave me an error MSG, "No such file or directory". The second one does give me about ten lines of codes with "" highlighted. I will have to study them closely. Since I am new to these, I am sure that I missed some subtle but critical steps for the first one. Could you please guild me through this?

  3. Thanks for bringing up the RFC6890 summary. Allow me to rephrase my project by saying that I am working on creating a very large new class of network that is "fully isolated" from both the current "public" and "private" networks. Thus, the choice of netblock for the routers on this new private (I call it semi-public) network is not restricted by any current Internet conventions. So, I would like to make OpenWrt as generic as possible. However, to avoid potential confusions or conflicts, even just to human minds, I am starting from the never explicitly used "Reserved for Future Use" netblock, 240/4.

Happy Holidays!

Abe (2018-12-21 12:19)

Dear Colleague,

I can tell you still haven't taken the time to learn the basics of OpenWrt nor that of modern networking. I would suggest and, in particular as well as a course on networking.

Your aspirations have been well-described in your previous posts, that have already been referred to in the search link above. The only confusion appears to be on your part. I am not at all confused by either your intended approach, nor that you have repeatedly made the error of utilizing this forum as a way of promoting a commercial endeavor that has no apparent benefit to the OpenWrt community. Searching your company's public materials reveal that it is now only a mailing address and that it has a history of trying to bring to market proprietary, closed networking solutions.

I am willing to provide specific, OpenWrt-related guidance consistent with your claim that you hold an MIT degree and the intelligence and skill set that I would typically associate with someone with those credentials. You will need to rise to that level.


For most routers it is defined at

You can simply use search at github:

1 Like
git clone
cd openwrt/
./scripts/feeds update -a
./scripts/feeds install -a
make download    # and accept the defaults (or adjust for your device) and write .config
egrep -rI '192\.168\.1\.1([^0-9]|$)' .

You may wish to complete a build as well, so that the various sources (such as the kernel) are extracted as they may have the restrictions on IP addresses that have been marked as Reserved that you keep asking about. Directions for building linked above.

1 Like

Likely yes; and you've have been told this multiple times in this community - in multiple threads. @jeff once again elaborated for you - as to why:


  • It is very important that you understand this for success in your task. Just because the default IP is, doesn't mean that you can merely change the IP to a Class E IP (e.g. This has been noted multiple times to you. This will likely require CODE REVIEW AND PROGRAMMING on your part, Your development team at Avinta Communications, Inc. should be able to do this for you.
  • Also be advised that is reserved and in Global use.

To be clear, RFC1112 states the IPs:

are reserved for future addressing modes.


1 Like

Also, to be clear -

  • it's again false that the IP space is for "future use"
  • this means that your proposed technology's addressing modes (i.e. Unicast, Broadcast and/or Anycast) in the Class E IP space doesn't fit into the REAL usage of the IPv4 space you wish to use.
  • This space's "IP behavior" is likely undefined and/or nonexistent in software code. Your development team at Avinta Communications, Inc. should be able to explain further.
  • You obviously know that your routers need to be clearly marked that they are incompatible with any other network or router on planet Earth.

Perhaps you could invent a new addressing mode instead?

(At this time I would suggest IPv6; but your paper expressing that it's a myth that: a.) IPv6 [fails to] fix end-to-end principle, b.) your belief you cannot firewall in IPv4 or IPv6 without NAT and c.) that a DMZ solves some of these in IPv4 - lead me to beleive it would be a fruitless endeavor.)

1 Like

Didn't you made a test yourself, and realized that Windows refuses to accept an IP address in that range from the DHCP server?

Isn't that proof enough that this is not going to work? Or did you manage to fix that?

1 Like

Hi, Jeff:

  1. Appreciate your instruction and information. However, I am perplexed by the continued off-track comments that kept on popping up, even though I purposely posted a generic technical question. Perhaps you are curious about what is behind me, but do not have a way to ask? Allow me to briefly address these topics from my perspective. Then, let's discuss them offline if you like (using eMail address at the end of my IETF Draft), so that we do not annoy the forum moderator again.

  2. When I was told to retake the one and the only one undergraduate course that I barely got a passing grade, I thought that I was doomed because it consisted of one quarter of my entire graduate course credits. It turned out to be my best performing course and actually helped me to land my first job! In reviewing it, I realized that instead of going through detailed derivations in the classroom as my undergraduate course was taught that bored me to death, the graduate class sessions were spent on interpreting Solid State Physics topics by daily life events, leaving the reading and derivation chores to homework. So, I was kept up in the evenings by the curiosity of how did it get to the end results. By the way, the professor who I am indebted forever was Mildred Dresselhaus, who is the "Queen of Carbon" fiber science. She was one of the students of Fermi who led the team making the first atomic bomb.

  3. Another tip of the iceberg topic to share: Please look for a sentence below

"Our protocols are not resilient enough and we are now seeing players break away to create their own protocols without necessarily seeking IETF permission."

about halfway down of the following APNIC blog. Can you think of any current happening that may be implied by this statement?

  1. Then, look for descriptions of RAN (Regional Area Network) in the following:

would you agree that there might be some correlation?

Happy Holidays!

Abe (2018-12-22 10:07)

Hi, hnyman:

Thanks. This gives me a lot to chew on.

Happy Holidays!

Abe (2018-12-22 10:09)

Hi, jeff:

Thanks for the additional instruction.

Abe (2018-12-22 10:10)

Hi, IIeachii:

  1. Perhaps you missed some of my earlier posts. They already hinted that the current OpenWrt is "unintentionally" routing packets with 240/4 netblock addresses.

  2. Allow me to submit the following latest lab notes:

Basically, columns AM - AP show that OpenWrt V18.06.1 is routing packets with 240/4 addresses sent and received by PCs loaded with Xubuntu V18.04.1

  1. However, preceding columns in the same spreadsheet indicate that the DHCP server itself in OpenWrt is not working with 240/4. This, my current question.

Happy Holidays!

Abe (2018-12-22 10:26)

Hi, lleachii:

  1. "Perhaps you could invent a new addressing mode instead?": Very close! Instead of "mode", if you use the word "space", we will be on the exactly same page. That is, the space that I am working on starts from ordinary "Private" network space, but using 240/4 netblock. So, it is isolated from the Internet "Public" space, but in between the two. I call it "Semi-Public/Private" space. Thus, the "new" router in this newly defined space is named SPR.

  2. Making sure that the interconnections among SPRs are not intertwined with the current "Public" and "Private" network spaces, this new space may grow significantly to be called a RAN (Regional Area Network) (between WAN and LANs). Basically, we could deploy many (number in the Billions) independent RANs, each is based on one IPv4 public address, and not much to do with current Internet rules and conventions.

Happy Holidays!

Abe (2018-12-22 10:43)

I've read that document before. I've missed nothing.

Perhaps YOU'RE missing something.

  • The behavior of Class E may be undefined. Not a general unicast behavior.
  • Then you're missing me. You are lying that 240/4 is reserved for future use. It's reserves for a future addressing mode. That means you cannot use normal IP routing methods already - for Class E space.

IANA clearly states they want a new addressing mode, nothing else.

Again, your technology does not exist. While I understand SPRs, Please use terms that exist currently. You then to look stupid relying on curricular logic instead. Thanks.

Please refrain from singing your posts.

1 Like

Hi, eduperez:

  1. You are correct, Windows based system was a no-go from the early tests. However, you might have missed some of the later developments on this forum that shed the light at the end of a long tunnel.

  2. Encouraged by one quick test from one colleague that his router appeared to be routing packets using 240/4 netblock, I went on to set up an OpenWrt router along with a couple of Xubuntu PCs. The results have been more positive than I had hoped. Please see the URL to a lab test notes at posting marker 9/15 of this thread.

  3. From the data, the OpenWrt is only one step away from fully supporting the routing of the 240/4 netblock. That is, if its DHCP server is set up to make use of the 240/4 netblock, OpenWrt would have expanded its capabilities to cover this new application.

Happy Holidays!

Abe (2018-12-22 11:02)


This makes sense!

But, the development team at Avinta Communications, Inc. needs to expand the capabilities, not OpenWrt!

DHCP is ran by multiple packages that can be installed on OpenWrt, dnsmasq is installed by default.

OpenWrt does not develop any of them!


This no longer seems to be an OpenWrt issue.

Happy Holidays! :star:


Are you offering a paid job at your company to implement this?

1 Like

Perhaps not. This could have been done years ago; but he's doing too well getting educated help for free.

  • He was told months ago that coding needed to be done.
  • I (and others in this community) proved to him that OpenWrt routed the arbitrary IP he wished to use in "his experiment" - he takes credit for that "discovery" alone (i.e. showing me a document I provided data for, and his pretending as if I've never seen it)
  • His company has a history of creating "black-box" products
  • He seems to use the least coding necessary to accomplish the task
  • It seem he's now seeking someone to explain where the IPs appear - so he can simply "change them."
  • When people try to explain the problem (e.g. that it's Unicast, Anycast and Broadcast in a range where it's not used nor programmed) - and that these code limitations may exist in multiple places:
    • he pretends the moderators are getting mad
    • he pretends he doesn't understand
    • and he pretends that no one understands him

Don't forget that:

  • He claims to hold "MS & EE degrees in EE from Massachusetts Institute of Technology"
  • He claims to be the VP Engineering at this company
  • Has claimed that he has multiple, qualified colleagues working on this
  • Has balked at requests that he perform tests himself
  • This company's presence is now nothing more than a mailing address in a strip mall
  • The company's website is broken in multiple ways and hasn't been updated for years
  • D&B / Hoovers indicates 5 employees and under US$800,000 gross sales (from perhaps when the company actually was licensing their technology)

There are far too many inconsistencies here to believe that there is much more sought than promotion of the continuing renewal of his draft through the strong link rating of the OpenWrt community, and perhaps someone to do the work.