Business to Business IP communication

[Glen Sykes]

I posted this subject on the Wainhouse Forum some time ago and the response was somewhat underwhelming.

This is something we are all (manufacturers, resellers and end-users alike) going to come to have to deal with.
As I see it at the moment, the only method of communicating between businesses is via the good old internet. However given the lack of standards for firewall and NAT traversal, and the absolute lack of control of the traffic, this is a real barrier.

It's (at least in the UK) only becoming realistic for a business to leverage more than 1 network provider to co-operate to provide end to end QoS (in fact I've encountered only 1 company thus far to provide the means to provide end to end QoS via direct peering between 2 large network providers). How far away is it that we can expect the providers to acknowledge that by co-operating in this manner that they will actually stimulate demand, rather than the insular attitude demonstrated thus far which just serves to prevent growth?

How does the VTCTalk Forum community see this developing?

[George]

I see that being a problem for some time to come. The fact is the change you are looking for would require an even deeper standardization of the internet than there already is.

It's really hard for a company to guaranteee solid QoS between 2 networks that they don't own. Then on top of that to guarantee that they can do it for any network having not even assessed the network makes it an even tougher challenge. So starting a company or division of a company based on a business model surrounding this would be a shaky venture indeed.

[Glen Sykes]

How do you see the likes of Masergy coming into this?

Also who will be responsible for the governance of a Global Dialling scheme to ensure E.164 numbers are not duplicated (will it be done in the same way as domain registration, or will the existing Telco's governing bodies control this?). We mustn't forget SIP either, and how this will change the landscape. Will it be SIP end to end eventually, or will it be SIP or H.323 provide the 'glue' between dissimilar networks? It's not beyond comprehension that the more mature H.323 will sit in the middle as the glue between organisations, with the (potentially) more feature rich SIP protocol within the boundaries of the organisation. It's probably still too early to speculate, but what do you all make of this?

Where QoS is concerned, I know of at least one company that has leveraged its relationship with 2 telco's in order to obtain end to end QoS (ie the telcos acknowledge the CoS and QoS tags, could this be the way forward?

My other quandary is secure boundary traversal. Given that there is no standard way of doing this, it's a bit of a minefield.

Finally, is this really a barrier, are businesses being held back by the lack of interconnectivity, or am I putting the cart before the horse (to coin a phrase!).

[tom9933]

On the dialing plan side you might check out the ViDe.Net site. They’ve already come up with a solution that’s been implemented in a lot of universities.
http://www.vide.net/

[Glen Sykes]

Hi Tom,

You make a good point, that is the Global Dialling Scheme that has been adopted across the world in education. In the UK, an organisation known as SuperJanet governs the dialling scheme and it works very well. It has also been extended from just universities to secondary and primary schools, as well as Libraries. It would make a lot of sense for network providers to incorporate this GDS so as to avoid duplication of numbers and ensure that educational establishments are not separated from the rest of the world by virtue of the numbering scheme.

The problem is who is going to make the network providers adhere to this GDS instead of creating thier own? I know of numerous H.323 installation that have completely proprietary numbering plans that would have to be completely re-written if they were to be integrated into this GDS.

The other problem is who is to say that numbering is the way to go anyway. Whilst it's true that we have to use numbers in order to ensure interoperability between H.323, H.320 and PSTN based devices, we also have another huge and mature addressing scheme in DNS.

[tom9933]

Glen,
I’m not sure how to make the service providers play ball, but the zone prefixing question may actually be a bit easier to answer. Maybe the move towards H.350 directories will help with this.
In our network the e.164 zone prefix is based on our actual phone numbers. So in ViDe we use the same zone prefix as the phone prefix for any phone on the university phone system. Then for each endpoint I use a 4 digit extension that in most cases matches either the user’s phone number or the phone number for the phone in that room. Now there are a few odd cases like conferences on the bridge, but those are using a block of addresses that have been assigned for special tasks by the phone guys. The other part of this is that the gatekeeper is setup to strip the zone prefix from calls that are received from the national gatekeeper (our neighbor). Some people choose to setup their networks like this and some just register all the endpoints with the full dialing string. I choose to implement striping at the gatekeeper so the local aliases would be easier to deal with.
On the H.323 alias (and maybe SIP down the road) side we use the email address structure. So all of the endpoints names match the email address of the user and in the case of a room system we use the building and room number as the name. Prefix stripping should in theory work the same way but at this time most people are only using the e.164 alias for out of zone calls.
Thoughts?

[Glen Sykes]

I think it would be much easier to implement a structure around DNS is the same way as E-mail than to use numbers, simply because this addressing scheme allows organisations to have control of their addressing within a domain. There is no comparable tier in E.164 to the domain at present (when looking at it from a point of view of PSTN, however in the education GDS there is), i.e. multiple organisations exist within an area code, but may well share the next 3 or 4 digits with other organisations to gain maximum use from the available digits, and it's up to the telco to assign these.

However, whilst DNS would appear to be the simplest addressing method to implement, it doesn't allow users (either H.323 or SIP) to be able to connect to either PSTN or ISDN based devices!

My thoughts are that an agreement between the global service providers needs to be arrived at on a structure of E.164 addresses that should be used. The GDS structure would make a lot of sense, given that it works and would ensure that we didn't end up with seperate, incompatible addressing schemes between education and the rest of the world. Finally an organisation needs to govern the use of E.164 prefixes, in the same way as DNS, where organisations are allocated say a 4 digit prefix within a particular area code, and allowed up to say 4 or 5 digits internally for extensions.