Sponsors
Sponsor Products
IPv4... it was a good ride!
posted by Greg Whynott  on Jan. 28, 2016, 11:30 a.m. (3 years, 24 days ago)
16 Responses     0 Plus One's     0 Comments  
Back in Apr 2015 I called ARIN and asked about the availability of netblocks. At that time I was told there were no more /22's which was the smallest blocks you could advertise to your upstream BGP peer at the time. I've no practical use for 1024 IPs then and now, but that is what it was if you wanted to have multiple paths. I was told 'we have no more /22's but there are 1000's of /24's, but they are going quick."..

1000's seemed like lots at the time...

Fast forward to Dec 2015 we get our multiple gig/isp connections installed. I start the BGP peering conversations with the ISPs, and apply for a /24. They are all gone! The only option is to re-apply to be put on a waiting list, I'd hate to guess at how long that would be.

so that all sucks. it sucks that a customer with multiple paths/ISPs can't have have their single IPv4/CIDR reachable by any ISP you peer with. We should of went full steam ahead on the IPv6 train back in the 90's... That is an option we can't seriously consider yet, still...

DNS RR is so not an good solution. its a duct tape fix... Funny all these challenges seem to be based on 20-40 year old protocols. Its time to upgrade all the internet core services to windows 10! naturally that is a joke, use unix. ;) but man the internet could use a firmware update.


Anyone have a /24 they want to sell? I can't accept RFC1918 networks, sorry the machine is broken.

-g





Thread Tags:
  discuss-at-studiosysadmins 

Response from Greg Whynott @ Feb. 7, 2019, 5 p.m.
I am beating myself up today... not sure how long ago but I did post something here a few years back.. I was looking for a /22 as that was the minimal allowed to advertise your network with BGP on the internet. Ended up getting one for free from ARIN but at the same time seen people selling /22's on an auction site the guy at ARIN pointed me to. At the time they were selling for about 3000 dollars. Was happy we got one for free.

Fast forward a year or so and another /22 requirement came up. This time ARIN wasn't able to help. Went to the auction site and seen they were selling for @6000-7000 dollars. I almost had to change my underwear. We ended up working with our ISP instead. ;) At that time I was talking with my friends and mused out loud "we should buy a bunch of /22's and sit on them"..

Today I log onto that site (https://www.ipv4auctions.com/) and see /22s are now going for $22 000 dollars! almost 4X the cost. would of been a good return on investment. They still have /18's for sale, if you have $330 000 kicking around... Hummm... that's 16 /22's, could make a quick $30 000.

The amount of summarized routes advertised on the internet is a bit staggering too. At that same time we were looking for /22's (6 or 7 years ago?) I was accepting the entire internet routing table on our external routers. It was at around 330 000 routes in the table. Today I see that is up to almost 770 000.. I suspect this is because it is 'ok' now to advertise a CIDR smaller than /22.

Julian, what is this BGP partitioning you are talking about? Link me please. ;) I never heard of it outside of how hardware was using partitioning to increase performance/scalability while dealing with a larger routing table. Is that the context you were speaking in?
psssst, wanna by a /24? its dankiest stuff you'll ever put down your pipe. :)
-g










On Fri, Jan 25, 2019 at 7:20 AM julian firminger <justdigitalfilm@gmail.com> wrote:
Sure, NAT at one level. At another BGP partitioning did.And the clouds... dont underestimate the swallowing of what were external services behind almost infinitely deep machine created subdomain structures. Back in the day, a service you provided needed an IP number. That just isnt true anymore.

Julian Firminger

Senior Systems AdministratorUnited Broadcast FacilitiesAmsterdam, The Netherlands


On Wed, Jan 23, 2019 at 11:29 PM greg whynott <greg.whynott@gmail.com> wrote:
NAT saved IPv4, its here for another generation or three I'd say.
-g

On Wed, Jan 23, 2019 at 5:57 AM Kevin Osbern <content@studiosysadmins.com> wrote:

I dont think we are in a position to force a change to IPv6 as of now. IPv4 doesnt seem to be exhausting ipv4 marketplace completely anytime soon. And some people are already changing to IPv6 and this will definitely increase the speed of IPv4.

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe
To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe
To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Julian Firminger @ Jan. 25, 2019, 7:25 a.m.
Sure, NAT at one level. At another BGP partitioning did.And the clouds... dont underestimate the swallowing of what were external services behind almost infinitely deep machine created subdomain structures. Back in the day, a service you provided needed an IP number. That just isnt true anymore.

Julian Firminger

Senior Systems AdministratorUnited Broadcast FacilitiesAmsterdam, The Netherlands


On Wed, Jan 23, 2019 at 11:29 PM greg whynott <greg.whynott@gmail.com> wrote:
NAT saved IPv4, its here for another generation or three I'd say.
-g

On Wed, Jan 23, 2019 at 5:57 AM Kevin Osbern <content@studiosysadmins.com> wrote:

I dont think we are in a position to force a change to IPv6 as of now. IPv4 doesnt seem to be exhausting ipv4 marketplace completely anytime soon. And some people are already changing to IPv6 and this will definitely increase the speed of IPv4.

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe
To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Greg Whynott @ Jan. 23, 2019, 5:30 p.m.
NAT saved IPv4, its here for another generation or three I'd say.
-g

On Wed, Jan 23, 2019 at 5:57 AM Kevin Osbern <content@studiosysadmins.com> wrote:

I dont think we are in a position to force a change to IPv6 as of now. IPv4 doesnt seem to be exhausting ipv4 marketplace completely anytime soon. And some people are already changing to IPv6 and this will definitely increase the speed of IPv4.

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Kevin Osbern @ Jan. 23, 2019, 5:57 a.m.

I don’t think we are in a position to force a change to IPv6 as of now. IPv4 doesn’t seem to be exhausting ipv4 marketplace completely anytime soon. And some people are already changing to IPv6 and this will definitely increase the speed of IPv4.


0 Plus One's     0 Comments  
   

Response from Greg Whynott @ Feb. 1, 2016, 11:15 a.m.
RFC1918 elements define 'private' networks. The ones Dan and yourself mention will never be found on today's internet routing tables. Therefore, for what the intention and scope was (Redundant paths with a common netblock behind them to be advertised and reachable via all paths), these will never work. Most BGP routers are configured to not accept RFC1918 advertisements, and the others will drop them on the floor.
The largest logical reason for not allowing the advertisement of RFC1918 networks is that "everyone" is using them, removing the 'uniqueness' of the address. Remember each IP/CIDR on the internet resolves to a unique node/network. We can't add routes to a 192.168/10.0/172.16 netowrks when there are millions of businesses using them. Everyone would be saying "i have the route to 192.168.0.0/16" and packets wouldn't have a chance at arriving where you intended them to.
hope that helps a bit. ;)
greg


On Mon, Feb 1, 2016 at 10:25 AM, Joachim Thuau <epac@korigan.net> wrote:
Just trying to point out that rfc1918 is but one set of private networks... Rather than offering it as an option.
The other thing is if you are advertising, is there anything actually preventing you from advertising rfc 1918 ranges along your other networks?(I legitimately don't know, but inquiring mind wants to find out...)
Thanks,Jok
On Feb 1, 2016, at 7:07 AM, greg whynott <greg.whynott@gmail.com> wrote:

Does your ISP take your RFC1918 networks and advertise them to the upstreams? for sure not so I'm not sure what you were getting at Joachim. help me out. ;)
greg

On Sat, Jan 30, 2016 at 8:10 PM, Joachim Thuau <epac@korigan.net> wrote:


On Jan 30, 2016, at 5:01 PM, Dan Mons <dmons@cuttingedge.com.au> wrote:

On 29 January 2016 at 02:25, greg whynott <greg.whynott@gmail.com> wrote:
Anyone have a /24 they want to sell? I can't accept RFC1918 networks,

Why not? I've happily run 40-location mutli-nationals with users in
the many thousands on RFC1918. What's the problem with it?


There are so many more options...
https://tools.ietf.org/html/rfc5735
Jok
To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe


0 Plus One's     0 Comments  
   

Response from Joachin Thuau @ Feb. 1, 2016, 10:30 a.m.
Just trying to point out that rfc1918 is but one set of private networks... Rather than offering it as an option.
The other thing is if you are advertising, is there anything actually preventing you from advertising rfc 1918 ranges along your other networks? (I legitimately don't know, but inquiring mind wants to find out...)
Thanks,Jok
On Feb 1, 2016, at 7:07 AM, greg whynott <greg.whynott@gmail.com> wrote:

Does your ISP take your RFC1918 networks and advertise them to the upstreams?  for sure not so I'm not sure what you were getting at Joachim.  help me out.  ;)
greg

On Sat, Jan 30, 2016 at 8:10 PM, Joachim Thuau <epac@korigan.net> wrote:


On Jan 30, 2016, at 5:01 PM, Dan Mons <dmons@cuttingedge.com.au> wrote:

On 29 January 2016 at 02:25, greg whynott <greg.whynott@gmail.com> wrote:
Anyone have a /24 they want to sell?   I can't accept RFC1918 networks,

Why not?  I've happily run 40-location mutli-nationals with users in
the many thousands on RFC1918.  What's the problem with it?


There are so many more options...
https://tools.ietf.org/html/rfc5735
Jok
To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Greg Whynott @ Feb. 1, 2016, 10:10 a.m.
Does your ISP take your RFC1918 networks and advertise them to the upstreams? for sure not so I'm not sure what you were getting at Joachim. help me out. ;)
greg

On Sat, Jan 30, 2016 at 8:10 PM, Joachim Thuau <epac@korigan.net> wrote:


On Jan 30, 2016, at 5:01 PM, Dan Mons <dmons@cuttingedge.com.au> wrote:

On 29 January 2016 at 02:25, greg whynott <greg.whynott@gmail.com> wrote:
Anyone have a /24 they want to sell? I can't accept RFC1918 networks,

Why not? I've happily run 40-location mutli-nationals with users in
the many thousands on RFC1918. What's the problem with it?


There are so many more options...
https://tools.ietf.org/html/rfc5735
Jok
To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe


0 Plus One's     0 Comments  
   

Response from Joachin Thuau @ Jan. 30, 2016, 8:15 p.m.


On Jan 30, 2016, at 5:01 PM, Dan Mons <dmons@cuttingedge.com.au> wrote:

On 29 January 2016 at 02:25, greg whynott <greg.whynott@gmail.com> wrote:
Anyone have a /24 they want to sell?   I can't accept RFC1918 networks,

Why not?  I've happily run 40-location mutli-nationals with users in
the many thousands on RFC1918.  What's the problem with it?


There are so many more options...
https://tools.ietf.org/html/rfc5735
Jok

0 Plus One's     0 Comments  
   

Response from Dan Mons @ Jan. 30, 2016, 8:05 p.m.
On 29 January 2016 at 02:25, greg whynott wrote: > Anyone have a /24 they want to sell? I can't accept RFC1918 networks, Why not? I've happily run 40-location mutli-nationals with users in the many thousands on RFC1918. What's the problem with it? -Dan To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Greg Whynott @ Jan. 29, 2016, 10:50 a.m.
wow 4 gigs... Time flys. The last time I took care of BGP was on a pair of 6509e's and we had 1 gig in them. One day I came into work and the internet was broken. The fix was to upgrade the 6509's to 2 gigs. Now 4. I wonder how much they would need if they started accepting CIDRs smaller than /24.. 16 GB of DDR7. ;)
On Thu, Jan 28, 2016 at 6:49 PM, Saker Klippsten <sakerk@gmail.com> wrote:
I think you need at least 8gb ram now. 2014 hit the 4gb mark.
http://blogs.cisco.com/sp/global-internet-routing-table-reaches-512k-milestone
We ended up installing with 16gb ram ;)

On Jan 28, 2016, at 5:37 PM, greg whynott <greg.whynott@gmail.com> wrote:

Same here Shane, when I worked at SGI/Alias our internal machines were numbered with real internet IPs, no RFC1918. I we (SGI/Cray/AW) owned not one but several /8's... I think my machines IP started with 144, that is going back a decade or so..... But there was talk of IPv6 and IPv4 depleation even when I was there in the 90's/2000's, I think. lol. memory is kaput....
The problem was created (or exaggerated) by ARIN and the like I think. There was a time where you couldn't advertise anything less than a /22. So today there are lots of companies out there with a /22 and may only have less than a /29 in use, just a few IPs for NAT, MX maybe DNS... With cloud services and the like the need to host internet services has gone away for many companies... I've been at or involved with a few BGP deployments now where this was the case. The one shop has a /22 and literately only 3 IPs assigned and in use, VPN, NAT/Outside interface and a web server that does nothing but redirect people landing on their old site since they moved everything to the cloud. With that said I understand memory was a big consideration and challange during the early days. I think most routers now have 2 gigs or more to hold the internet table, and now that they are accepting /24's, that has to be even higher by now(?).. 3 out of 1024 isn't bad, right? ;)
One might speculate less than 1/2 of the potential allocation space is actually in use. i'm sure there is a site with that info...




On Thu, Jan 28, 2016 at 4:58 PM, Shane McEwan <shane@mcewan.id.au> wrote:
On 28/01/16 16:25, greg whynott wrote:
We should of went full steam ahead on the IPv6 train back in
the 90's...

Except back in the '90s this never seemed like a problem.

At my very first job back in 1990 we had a Class B address range (/16 as the kids would call it these days. 140.79.0.0 if I remember correctly). 65536 IP addresses for 150 staff . . . most of whom didn't even have a computer. Good times . . .

Shane.

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe


0 Plus One's     0 Comments  
   

Response from Saker Klippsten @ Jan. 28, 2016, 6:50 p.m.
I think you need at least 8gb ram now. 2014 hit the 4gb mark. 
http://blogs.cisco.com/sp/global-internet-routing-table-reaches-512k-milestone
We ended up installing with 16gb ram ;)

On Jan 28, 2016, at 5:37 PM, greg whynott <greg.whynott@gmail.com> wrote:

Same here Shane,  when I worked at SGI/Alias our internal machines were numbered with real internet IPs,  no RFC1918.   I we (SGI/Cray/AW) owned not one but several /8's...  I think my machines IP started with 144,   that is going back a decade or so.....  But there was talk of IPv6 and IPv4 depleation even when I was there in the 90's/2000's,  I think.  lol. memory is kaput.... 
The problem was created (or exaggerated) by ARIN and the like I think.   There was a time where you couldn't advertise anything less than a /22.  So today there are lots of companies out there with a /22 and may only have less than a /29 in use,  just a few IPs for NAT, MX maybe DNS...    With cloud services and the like the need to host internet services has gone away for many companies...  I've been at or involved with a few BGP deployments now where this was the case.    The one shop has a /22 and literately only 3 IPs assigned and in use,  VPN,  NAT/Outside interface and a web server that does nothing but redirect people landing on their old site since they moved everything to the cloud.   With that said I understand memory was a big consideration and challange during the early days.   I think most routers now have 2 gigs or more to hold the internet table,  and now that they are accepting /24's,  that has to be even higher by now(?)..  3 out of 1024 isn't bad,  right?  ;)
One might speculate less than 1/2 of the potential allocation space is actually in use.    i'm sure there is a site with that info...




On Thu, Jan 28, 2016 at 4:58 PM, Shane McEwan <shane@mcewan.id.au> wrote:
On 28/01/16 16:25, greg whynott wrote:
We should of went full steam ahead on the IPv6 train back in
the 90's...

Except back in the '90s this never seemed like a problem.

At my very first job back in 1990 we had a Class B address range (/16 as the kids would call it these days. 140.79.0.0 if I remember correctly). 65536 IP addresses for 150 staff . . . most of whom didn't even have a computer. Good times . . .

Shane.

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Greg Whynott @ Jan. 28, 2016, 5:40 p.m.
Same here Shane, when I worked at SGI/Alias our internal machines were numbered with real internet IPs, no RFC1918. I we (SGI/Cray/AW) owned not one but several /8's... I think my machines IP started with 144, that is going back a decade or so..... But there was talk of IPv6 and IPv4 depleation even when I was there in the 90's/2000's, I think. lol. memory is kaput....
The problem was created (or exaggerated) by ARIN and the like I think. There was a time where you couldn't advertise anything less than a /22. So today there are lots of companies out there with a /22 and may only have less than a /29 in use, just a few IPs for NAT, MX maybe DNS... With cloud services and the like the need to host internet services has gone away for many companies... I've been at or involved with a few BGP deployments now where this was the case. The one shop has a /22 and literately only 3 IPs assigned and in use, VPN, NAT/Outside interface and a web server that does nothing but redirect people landing on their old site since they moved everything to the cloud. With that said I understand memory was a big consideration and challange during the early days. I think most routers now have 2 gigs or more to hold the internet table, and now that they are accepting /24's, that has to be even higher by now(?).. 3 out of 1024 isn't bad, right? ;)
One might speculate less than 1/2 of the potential allocation space is actually in use. i'm sure there is a site with that info...




On Thu, Jan 28, 2016 at 4:58 PM, Shane McEwan <shane@mcewan.id.au> wrote:
On 28/01/16 16:25, greg whynott wrote:
We should of went full steam ahead on the IPv6 train back in
the 90's...

Except back in the '90s this never seemed like a problem.

At my very first job back in 1990 we had a Class B address range (/16 as the kids would call it these days. 140.79.0.0 if I remember correctly). 65536 IP addresses for 150 staff . . . most of whom didn't even have a computer. Good times . . .

Shane.

To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe


0 Plus One's     0 Comments  
   

Response from Shane McEwan @ Jan. 28, 2016, 5 p.m.
On 28/01/16 16:25, greg whynott wrote: > We should of went full steam ahead on the IPv6 train back in > the 90's... Except back in the '90s this never seemed like a problem. At my very first job back in 1990 we had a Class B address range (/16 as the kids would call it these days. 140.79.0.0 if I remember correctly). 65536 IP addresses for 150 staff . . . most of whom didn't even have a computer. Good times . . . Shane. To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Greg Whynott @ Jan. 28, 2016, 2:30 p.m.
I discovered www.ipv4auctions.com 9000 dollars for a /22, 3000 for a /24, which doesn't really seem like a lot of money for what a business gets, if they have that need. still quite a bit when it should be and was free.
Impossible and just a dream - if you are not using IPv4 space, it should be mandatory to have it returned.

just like this new rule the city imposed this year with garbage.... well sorta... they returned my garbage the other day...
I put out my trash can on wheels (new program they have, so they don't even get out of the truck now, the 'machine' does.. ( another win for the unions.... *sigh*). Anyway came home last night and noticed they didn't take my garbage. I mused about this out loud and my 11 year old daughter informed me that they don't take them unless it is full (mine was 1/4 full).
1. funny how she knew and I didn't...2. I have the 'large' size container (ordered it when there was 5 people living in the house, now there is just me), so if I got a small size that would of been 3/4 full. so whats the deal? why do they base the ratio to the container size and not on a wieght or some other metric? why do they care? you are on my street at my garbage can. does it save you money not taking it after you determined it doesn't meet some acceptable level of garbage amount policy? f'ing useless unions. lol...
-g







On Thu, Jan 28, 2016 at 12:57 PM, Saker Klippsten <sakerk@gmail.com> wrote:
Not to add insult to injury...We were one of the folks who got in under the gun. 4 days before they ran out. Had filled out the back and forth paper work. Needed one last sign off. I was lazy about and then saw they were down to 100 /24 on their site Got ours in Nov.

Our Cisco ASRs just showed up :)

-s



> On Jan 28, 2016, at 11:25 AM, greg whynott <greg.whynott@gmail.com> wrote:
>
> Back in Apr 2015 I called ARIN and asked about the availability of netblocks. At that time I was told there were no more /22's which was the smallest blocks you could advertise to your upstream BGP peer at the time. I've no practical use for 1024 IPs then and now, but that is what it was if you wanted to have multiple paths. I was told 'we have no more /22's but there are 1000's of /24's, but they are going quick."..
>
> 1000's seemed like lots at the time...
>
> Fast forward to Dec 2015 we get our multiple gig/isp connections installed. I start the BGP peering conversations with the ISPs, and apply for a /24. They are all gone! The only option is to re-apply to be put on a waiting list, I'd hate to guess at how long that would be.
>
> so that all sucks. it sucks that a customer with multiple paths/ISPs can't have have their single IPv4/CIDR reachable by any ISP you peer with. We should of went full steam ahead on the IPv6 train back in the 90's... That is an option we can't seriously consider yet, still...
>
> DNS RR is so not an good solution. its a duct tape fix... Funny all these challenges seem to be based on 20-40 year old protocols. Its time to upgrade all the internet core services to windows 10! naturally that is a joke, use unix. ;) but man the internet could use a firmware update.
>
>
> Anyone have a /24 they want to sell? I can't accept RFC1918 networks, sorry the machine is broken.
>
> -g
>
>
>
>
> To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe
To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe


0 Plus One's     0 Comments  
   

Response from Saker Klippsten @ Jan. 28, 2016, 1 p.m.
Not to add insult to injury...We were one of the folks who got in under the gun. 4 days before they ran out. Had filled out the back and forth paper work. Needed one last sign off. I was lazy about and then saw they were down to 100 /24 on their site Got ours in Nov. Our Cisco ASRs just showed up :) -s > On Jan 28, 2016, at 11:25 AM, greg whynott wrote: > > Back in Apr 2015 I called ARIN and asked about the availability of netblocks. At that time I was told there were no more /22's which was the smallest blocks you could advertise to your upstream BGP peer at the time. I've no practical use for 1024 IPs then and now, but that is what it was if you wanted to have multiple paths. I was told 'we have no more /22's but there are 1000's of /24's, but they are going quick.".. > > 1000's seemed like lots at the time... > > Fast forward to Dec 2015 we get our multiple gig/isp connections installed. I start the BGP peering conversations with the ISPs, and apply for a /24. They are all gone! The only option is to re-apply to be put on a waiting list, I'd hate to guess at how long that would be. > > so that all sucks. it sucks that a customer with multiple paths/ISPs can't have have their single IPv4/CIDR reachable by any ISP you peer with. We should of went full steam ahead on the IPv6 train back in the 90's... That is an option we can't seriously consider yet, still... > > DNS RR is so not an good solution. its a duct tape fix... Funny all these challenges seem to be based on 20-40 year old protocols. Its time to upgrade all the internet core services to windows 10! naturally that is a joke, use unix. ;) but man the internet could use a firmware update. > > > Anyone have a /24 they want to sell? I can't accept RFC1918 networks, sorry the machine is broken. > > -g > > > > > To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe

0 Plus One's     0 Comments  
   

Response from Craig Dibble @ Jan. 28, 2016, 12:10 p.m.
Yikes!

There was a roaring trade in Romania, and I seem to recall some African countries were jumping on that particular bandwagon recently.

Good luck!

Craig

On 28 January 2016 at 16:25, greg whynott <greg.whynott@gmail.com> wrote:
Back in Apr 2015 I called ARIN and asked about the availability of netblocks. At that time I was told there were no more /22's which was the smallest blocks you could advertise to your upstream BGP peer at the time. I've no practical use for 1024 IPs then and now, but that is what it was if you wanted to have multiple paths. I was told 'we have no more /22's but there are 1000's of /24's, but they are going quick."..

1000's seemed like lots at the time...

Fast forward to Dec 2015 we get our multiple gig/isp connections installed. I start the BGP peering conversations with the ISPs, and apply for a /24. They are all gone! The only option is to re-apply to be put on a waiting list, I'd hate to guess at how long that would be.

so that all sucks. it sucks that a customer with multiple paths/ISPs can't have have their single IPv4/CIDR reachable by any ISP you peer with. We should of went full steam ahead on the IPv6 train back in the 90's... That is an option we can't seriously consider yet, still...

DNS RR is so not an good solution. its a duct tape fix... Funny all these challenges seem to be based on 20-40 year old protocols. Its time to upgrade all the internet core services to windows 10! naturally that is a joke, use unix. ;) but man the internet could use a firmware update.


Anyone have a /24 they want to sell? I can't accept RFC1918 networks, sorry the machine is broken.

-g





To unsubscribe from the list send a blank e-mail to mailto:studiosysadmins-discuss-request@studiosysadmins.com?subject=unsubscribe


0 Plus One's     0 Comments