Third party open access is the single most significant issue facing cable and its high speed Internet customers today.
I define third party open access when the subscriber's data packets travel to the servers and through the network of whichever third party provider they are using. For instance, if my third party provider were PSINet Canada, my packets would travel directly to and through PSINet Canada's network first and then through the rest of the Internet.
In this case, neither @Home nor any cable MSO would have any hand in providing transport for the subscriber's packets to their ultimate destination. The MSO/infrastructure provider would only be responsible for enabling the subscriber to receive broadband internet access and allowing their packets to be transported directly to the provider of their choice.
I make this specific definition in order to rule out service resale. Service resale is not real third party open access, as the subscriber's packets will still be routed through the incumbent backbone provider's network first before they are able to reach their destination. For example, I may use Sprint Canada as my resale provider, but my packets will still travel over the @Home network in order to reach any external destinations. This, in my opinion, negates any of the significant benefits that legitimate open access provides for the consumer.
From a consumer perspective, open access should be simple, scalable, tolerant to change and unobtrusive. From a provider perspective, ease of deployment, simplicity, scalability, efficient network management, and effective billing are the important factors. There is some definite overlap in what subscribers and providers view as important on the open access front.
What methods, then, can be implemented to provision third party open access? There are three primary ways of doing this: IP tunneling (this includes something called PPPoE, otherwise known as Point-to-Point Protocol over Ethernet), VLAN (Virtual Local Area Network) encapsulation, and source network address routing, a type of policy-based routing.
What's the best way of provisioning open access? The superior method should most appropriately fulfill the needs of providers and consumers alike. This rules out PPP over Ethernet, as it has proven to be obtrusive, complicated (due to it being completely non-standard and non-tested), resistant to future change (again, as it is non-standard) and not very scalable.
From an engineering and logical standpoint, running PPPoE makes no sense whatsoever and serves no purpose for the consumer. It provides no security features, no performance premium, no expedition of deployment and many existing users who have experienced it are heartily voicing their disapproval for the technology. In fact, over 1,700 subscribers to Bell Sympatico's high speed edition Internet service, which employs Nortel's megabit modem DSL technology, have signed a petition against PPPoE on the Sympatico Users' Group website (http://sympaticousers.org/). Many of those signatures were provided before the service passed the 35,000 subscriber mark, making a whopping 5% of the user base unhappy with PPPoE.
The truth of the matter is that PPPoE and IP tunneling in general is a very poor method for provisioning third party open access to any broadband infrastructure. It is definitely the most inferior of the three methods named above.
VLAN encapsulation and policy-based routing solutions are highly preferable to any kind of IP tunneling, as they don't require the subscriber to install software on their computer and the whole process is completely transparent to the user. In addition, open access is provisioned more quickly, easily and effectively with greater simplicity, and at lesser cost, simply because the provider does not have to rely on each individual software installation to be performed correctly -- because there is nothing to be installed. This greatly reduces technical support costs and expedites the number of problem-free installations that can be performed within any given period of time.
I believe that if Bell does not drop PPPoE, the other two prevailing PPPoE-free broadband alternatives -- cable and microwave -- will completely eclipse DSL as the future of high-speed internet access. PPP over Ethernet sets internet access technology back at least five years so I implore cable and microwave providers not to follow Bell's path.
*****
I started out like every one of my peers -- as a subscriber to the Rogers@Home cable modem service since October 1998. After six weeks, however, the entire service collapsed. That's not hyperbole; it's the truth. The problems were rectified a few weeks later, but resurfaced again over the coming months.
The most significant problem was that there was no communication between Rogers and its high-speed customers. Technical support, the only line of communication that normally exists between provider and subscriber, was either completely unavailable (telephone hold times exceeded 60 minutes on average; e-mails went totally unanswered), or unable to provide any helpful information when available. In fact, no fault was admitted by Rogers for any of the resultant network problems and that is a primary reason why the Rogers @Home Users Association (RHUA) was formed.
Shortly after our organization was born, communication with management was established, press attention was garnered and things improved a number of months later. That isn't to say that our service is devoid of problems nowadays, but we're better off now than we were a year ago.
This is the first of what will be a series of columns. The purpose of my writings is to offer insight on consumer issues and topics of mutual importance to the provider, the consumer and the regulator as they relate to high speed Internet access.
Christopher Weisdorf is an independent telecommunications consultant and the technical director of the RHUA. He can be contacted at chrisw@rbua.org. The RHUA is an independent association of @Home Canada customers.