Hi Robert,

If I remember correctly we agreed to set the costs for "private" <-> "private" to 0. The definitions in section 2.1 and 2.2 as well as the appendices A.2 and A.5 say so too.

So Wendy's server response seems correct to me. I also checked the BENOCS server which returned the correct cost value to me.

ECS Test 2 is indeed not correct.

Thanks
Hans




On 17.09.2015 13:33, Chenguohai wrote:

Hi Wendy and Hans,

I found an error about section 4.4.2 in ALTO interop document [http://tools.ietf.org/html/draft-roome-alto-interop-ietf93-01]. Section 4.4.2 indicates there is no costs in response but appendix A.2. defines a 'routingcost' cost from *private* to *private*.


Did I misunderstand this. I made a record about this and we may update in next submitting if my understanding is right.

@Wendy your server does return a cost (ipv4:10.0.1.0 -- > ipv4:10.0.1.1 ).
http://alto.benocs.berlin:8000/directory does not return any cost.


-------------------------------------------------
4.4.2.  ECS Test 2

   This test determines the costs between endpoints in the "default"
   PID:

      Query:
        sources:
              ipv4:10.0.1.0 ("private")
              ipv6:::2      ("default")
        destinations:
              ipv4:10.0.1.1 ("private")
              ipv6:::1:2    ("default")

                           Figure 15: ECS Test 2

Since no costs are defined between those PIDs the server MUST return
   an ECS response containing no costs.

------------------------------------------------

*A.2*<http://tools.ietf.org/html/draft-roome-alto-interop-ietf93-01#appendix-A.2>*.Default "routingcost" Cost Map*

"cost-map": {

"default": {

"mine": 60.0,"mine1": 63.0,"mine2": 64.0,"mine1a": 65.0,

"mine3": 65.0 },

"linklocal": {

"linklocal": 0.0 },

"loopback": {

"loopback": 0.0 },

"mine": {

"mine": 0.0,"mine1": 3.0,"mine2": 4.0,"mine1a": 5.0,

"mine3": 5.0,"peer1": 20.0,"peer2": 25.0,"tran1": 35.0,

"tran2": 45.0,"default": 60.0 },

"mine1": {

"mine": 3.0,"mine1": 0.0,"mine2": 6.5,"mine1a": 2.0,

"mine3": 8.0,"peer1": 23.0,"peer2": 28.0,"tran1": 38.0,

"tran2": 48.0,"default": 63.0 },

"mine1a": {

"mine": 5.0,"mine1": 2.0,"mine2": 4.5,"mine1a": 0.0,

"mine3": 10.0,"peer1": 25.0,"peer2": 30.0,"tran1": 40.0,

"tran2": 50.0,"default": 65.0 },

"mine2": {

"mine": 4.0,"mine1": 7.0,"mine2": 0.0,"mine1a": 9.0,

"mine3": 9.0,"peer1": 24.0,"peer2": 29.0,"tran1": 39.0,

"tran2": 49.0,"default": 64.0 },

"mine3": {

"mine": 5.0,"mine1": 8.0,"mine2": 9.0,"mine1a": 10.0,

"mine3": 0.0,"peer1": 25.0,"peer2": 30.0,"tran1": 40.0,

"tran2": 50.0,"default": 65.0 },

"peer1": {

"mine": 20.0,"mine1": 23.0,"mine2": 24.0,"mine1a": 25.0,

"mine3": 25.0,"peer1": 0.0

},

"peer2": {

"mine": 25.0,"mine1": 28.0,"mine2": 29.0,"mine1a": 30.0,

"mine3": 30.0,"peer2": 0.0 },

"private": {

"private": 0.0 },

BR

G.RobertChen

*From:*alto [mailto:[email protected]] *On Behalf Of *ChenGuohai
*Sent:* Thursday, September 17, 2015 6:12 PM
*To:* Roome Wendy; Seidel Hans
*Cc:* [email protected]
*Subject:* Re: [alto] On-line interop-93 test client

------------------------------------------------------------------------

Date: Fri, 4 Sep 2015 15:04:51 -0400
Subject: Re: [alto] On-line interop-93 test client
From: [email protected]<mailto:[email protected]>
To: [email protected]<mailto:[email protected]>
CC: [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]>

Hans,

I have revised the interop document. See attached.

NOTE: Please check the link costs in Figures 5 & 10. I did *not* update them from the version you gave me last week. In particular I am sure that Figure 5 is out-of-date, since that uses your original link costs.

I dropped the ECS tests for omitted source & destination. Those were always flakey, because the client address is unpredictable. Let me know if you think those are worth putting back in.

- Wendy Roome

*From: *Hans Seidel <<mailto:[email protected]>[email protected]>
*Date: *Thu, September 3, 2015 at 09:41
*To: *Wendy Roome <<mailto:[email protected]>[email protected]> *Cc: *"<mailto:[email protected]>[email protected]" <<mailto:[email protected]>[email protected]>, ChenGuohai <<mailto:[email protected]>[email protected]>
*Subject: *Re: [alto] On-line interop-93 test client

Wendy

Server is updated and running. Please check if the results are now as expected.

Thanks
Hans

On 02.09.2015 19:43, Wendy Roome wrote:

    Hans,

    Okay. Let me know when you have revised your server.

    - Wendy

    *From: *Hans Seidel <<mailto:[email protected]>[email protected]>
    *Date: *Wed, September 2, 2015 at 12:08
    *To: *Wendy Roome
    <<mailto:[email protected]>[email protected]>
    *Cc: *"<mailto:[email protected]>[email protected]"
    <<mailto:[email protected]>[email protected]>, ChenGuohai
    <[email protected]<mailto:[email protected]>>
    *Subject: *Re: [alto] On-line interop-93 test client

    Wendy,

    Thank you. No I got it.

    Yes. At the moment we rank the costs for each source separately.
    This also concerns endpoint cost service which I have to fix
    accordingly as well.

    Hans


_______________________________________________
alto mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/alto

Reply via email to