SNASw set up questions.

[a.mountouris]

Hello Forum, could you please give your guidance on the following issues?

A/. Can an upstream interface be used also as a downstream for locally attached PUs? I mean, we have a SNASw (HPR/IP actually) configured router which has an ethernet interface towards the mainframe that is the upstream interface. On the same LAN we have also some PCs. Can we use as destination MAC of these PCs the MAC of the router's ethernet interface instead the mainframe's OSA MAC??

B/. Case of remote SNASw (HPR/IP) router. In this case, we have configured the HPRIP port tied with the loopback and the link destination is the IP of the mainframe, so we have established an EE connection between the router and the mainframe (port and link actibe). Do we have to code also in the snasw process the ethernet interface that carries the downstream network?

[htranphu · Accepted solution]

A/ Yes.

A SNASw port (~ an interface) can be used both for defining an upstream link, and for receiving connections from downstream devices.

B/ Yes, you still need to configure a SNASw port for the downstream interface.

Regards,

tran

[romney]

Hi Apostolos,

to further clarify ...

A.

snasw port EE hpr-ip Ethernet 1

snasw port DWNSTRM Ethernet 1 conntype nohpr

snasw link HOST port EE ip-dest 10.10.10.1

B.

snasw port EE hpr-ip Loopback 0

snasw port DWNSTRM Ethernet 1 conntype nohpr

snasw link HOST port EE ip-dest 10.10.10.1

There are some restrictions - you cannot have more that one hpr-ip port on the same interface. i.e.

snasw port EE1 hpr-ip Loopback 0

snasw port EE2 hpr-ip Loopback 0

is not allowed.

You also cannot have more than one non hpr-ip port on the same interface unless you specify different saps. i.e.

snasw port DWNSTRM1 Ethernet 1 conntype nohpr

snasw port DWNSTRM2 Ethernet 1 conntype nohpr

is not allowed, but

snasw port DWNSTRM1 Ethernet 1 conntype nohpr

snasw port DWNSTRM2 Ethernet 2 conntype nohpr sap 8

is allowed (but downstream devices must specify a target sap of 8 rather than the default of 4 in order to connect to the second port).

- Ray

[a.mountouris]

Hi Ray,

I did't see this update when I sent my previous comment requesting a clarifcation. So, I can use at the same time the Ethernet1 both for upstream and downstream connections. Is it mandatory to code the DWNSTRM port for the downstream connections or you just put it there to highlight the example?

As I understood from the second example (B) it is mandatory to define the downstream port.

Regards, Apostolos.

[romney]

Apostolos,

in your example A the devices on the LAN could only use the EE port if they are capable of connecting via hpr-ip, which is why I showed the addition of a non-EE port, DWNSTRM. If they are capable of starting an hpr-ip link then they could use the EE port, but snaswitch would play a limited role (it would be an intermediate HPR ANR node and could not act as DLUR for those devices - dependent LUs are not supported over hpr-ip links). Can you tell me more about the devices in example A?

- Ray

[a.mountouris]

Devices in example A are PCs running PCOM emulator (5.1 I think). However we noticed a peculiar behaviour.A test PC on the same LAN as the ethernet interface that is the upstream link to the host (EE connction), established connection with the mainframe using as destination MAC the MAC of a test vdlc interface! The SNASw router also runs DLSw for the remote branches, but we din't expect that the local PCs ould use the vdlc MAC since they are not in the dlsw cloud.

I must say I do not fully understand how the SNASw ad DLSw processes interoperate with each other.

[mbinzer]

Hi,

in most of the cases the vdlc mac address will work but it is still not the supported/recommended way to use.

In your case, if you need the same mac address for the lan users than for those coming in over dlsw, you can simply configure the same mac address on the ethernet interface and define the ethernet interface as snasw port.

This way the pc will connect through the ethernet snasw port and not vdlc.

As i have tried to explain earlier vdlc is specifically designed to allow a connection between dlsw and snasw/appn, dspu. I think that are all of the users of vdlc. Appn is not an option anymore since snasw has replaced pure appn in cisco ios.

Dlsw historically was developed to provide bridging between two lan segments separated by a arbitrary tcp/ip network.

In this scenario a lan attached client on the branch end would connect via dlsw to a snasw router on the lan at the host end, two routers. one doing dlsw the other one snasw, could also be a FEP.

Vdlc, virtual dlc, basicaly represents the lan on the host end and you can thus connect directly to snasw in the same router.

Bottom line is that the connection through a lan port to VDLC is not supported. You need to define either the physical lan interface as snasw port or use hsrp.

thanks...

Matthias

[a.mountouris]

Hello Matthias and thanks for the clarifications. I have one more question about VDLC. Can I define more than one VDLCs on the same router? For example can I have two VDLC interfaces (and consequently code two source route bridge groups)?

Can I use a MAC address of the form 4200........? In an IBM manual about coding MACs on an OSA interface, the 42...... form is an ethernet MAC address. Does this apply to VDLC also?

[mbinzer]

Hi,

yes you can have multiple vdlc's on the same router.

Yes you can also have multiple source-bridge ring-groups on the same router. Dlsw will by default talk to all virtual ring-groups defined on the router.

You can also have multiple snasw vdlc ports defined with different mac addresses on the same virtual ring-group.

In respect to the mac address format of the vdlc port.

Dlsw internal sees all mac addresses in tokenring format. The format of the mac address configured on the snasw port is "token ring format", non canonical. So if you configure i.e. 4000.3745.0000 as your vdlc mac address on a snasw port than you will have to configure your clients, if they are on a remot ethernet behind dlsw, to connect to 0200.eca2.0000.

If the clients are on a tokerning behind a remot cisco router than you would configure their dmac always as 4000.3745.0000.

In respect to the dlsw/vdlc/snasw router the mac address format doesnt really matter since the downstream session is terminated in this router. It is not going anywhwere outside of the router. Only the upstream session is going out of the router again and this is done via EE hpr/ip.

So for practical purposes you always configure the snasw vdlc port mac address to match whatever is already configured as the dmac on your end systems. Just make sure you keep the bitswapping in mind in case you have to convert the media. I.e as i have already stated when the clients are on ethernet.

For redundancy you would simply configure the same snasw vdlc port mac address on two routers. Your remote router has a dlsw peer to both of them and learns the same mac address via both peers. Off course both dlsw/snasw routers at the head end must have upstream EE hpr/ip connectivity to the same resource/host.

Let me know if this answers your questions.

thanks...

Matthias

[a.mountouris]

Matthias thank you very much for the answers, they are clear and to the point.

I have just tested it (3 vdlcs under the same ring) and it works. The remarkable thing is that all tests worked regardless the media type I coded or MAC address. In the PCOM configuration, you indicate a destination MAC and you have to specify either ethernet or t/r type. For vdlcs, both configurations worked!!!

And another thing is that for one of our MACs, the bitswaped is the same!! Check the tool for 420042004200.

Anyway, the test worked and I am most grateful for your help.

P.S. Now that we can code multiple destination MACs at headquarter site (as vdlcs) there is another issue that came out. In a central branch we have two routers to carry the dlsw to the headquarter. So I guess we have to code ethernet redunduncy feature between them, am I right?

[mbinzer · Accepted solution]

Apostolos,

yes, if you have 2 dlsw routers sitting on the same ethernet segment, doing dlsw to the same remote resources ( doenst matter if this is to vdlc or i.e FEP's ) you must use dlsw ethernet redundancy between those two routers to be sure you keep dlsw and the ethernet switch happy.

thanks...

Matthias

[a.mountouris]

Thanks tran,

just a clarification about A. I mean at the same time the port being upstream and downstream, is it still possible?

And another thing about VDLC. As I understood, the VDLC concept is valid in case we have DLSw and SNASw, so SNASw can be imported into DLSw cloud. However, if we have EE connection between the Mainframe and all the way down to the branch office router (meaning no DLSw at all), then we can't really use VDLC interface in that router as the destination MAC of the branch PUs, am I right?

Regards, Apostolos.

[mbinzer]

Hi,

i let Tran clearify the first part of the question since i am not the expert in this area.

For the second part of your question. VDLC is only usable between dlsw and snasw in your setup. If you dont use dlsw, then you can not use VDLC.

For the EE hpr/ip upstream connection you certainly dont need VDLC at all.

For the downstream connections you can use the physical interface as snasw port, this is the most simple case.

If your medium would be a tokenring you use virtual tokenrings to get redundancy, by using source bridging from the physcial tokenring to the virtual tokenring interfaces which then serve as snasw port.

You can then have two routers on the same ring using the same mac address on a virtual tokenring and have two different srb path into those routers.

If you use ethernet interfaces you can use hsrp between two routers to get redundancy for snasw aswell. The dmac for the downstream clients will then be configured as hsrp mac address and will only be active on the router which hsrp active at the time.

VDLC is only a concept to be used when a downstream client connects into snasw via dlsw to have the ability to connect inside the physicaly same router to snasw.

thanks...

Matthias

[htranphu · Accepted solution]

Apostolos,

Regarding the config A, you can use the same hpr-ip port for upstream and downstream connections (assuming that the downstream devices do support hpr-ip as Ray has indicated); it will be IP between SNASw and the downstream device, which will point to the IP address of the interface of the SNASw port (no MAC address involved here).

As Matthias has explained, if you put SNASw on the branch office router, the downstream devices could probably be directly connected to SNASw via either an Ethernet port or an HPR-IP port (if the downstream device supports), and there would be no need for any VDLC port.

Regards,

tran