When you sign up for DSL, the speeds are "up to". This means you get what you get. If you have long lines or terrible wiring or whatever that slows down your 3.0/384 to only give you 700/100, well thats life. If you have "full speed" and they rewire a trunk to your office and the speed drops to 1/4 of that, well that's life. (Been there.)
With T1, BRI, PRI, etc... you get what you sign up for. If sign up for
1.5/1.5 then they give 1.5/1.5 at your demarc. If they have to add repeaters, run new wire, come out and tinker with things after a storm, whatever, you get the agreed rate.
In addition, a T-1 will have service response timed, sometimes measured in minutes. Your home DSL can go down for prolonged periods with no response required.
It all comes down to the "grade of service" you are purchasing from your supplier.
At the top end of the "grades" are the "Commercial Services" such as a T1 type Data Circuit. It is a dedicated Point-to-Point service, and the bandwidth is 100% guaranteed bandwidth and quality of service, delivered between 2 well defined end-points. NOTHING in the path your data takes is shared by anyone else, it's 100% guaranteed to be all yours. In addition, most of these types of services are actively monitored by your provider and they often respond to issues they notice before you are even are aware of them (while this is highly desirable it is not always reality.....;-)). In this part of the world (New Zealand) it is also possible to purchase a "slightly cheaper" Point-to-Point Data Circuit that is NOT actively monitored, however the provider tries to discourage these as their "incident" and support costs can be quite high. These types are Data Circutis are usually about 10% - 20% cheaper than the full T1 equivalent cost, so they are still not cheap bandwidth.
At the bottom end of the "grades" are "Consumer Services" such as PSTN (Dial-up), Cable and ADSL type technologies. These cannnot provide this same level of service guarantee, as it does nt have 2 end-poits and the PATH data takes is HSARED at some point, so there canont be guaranteed bandwidth, and cannot be guaranteed "end to end" delivery type of service similar to a full T1 Data Circuit. Due to the shared nature of the data path, you cannot be offereed anywhere near the same guaranteed quality of service.
So bandwidth alone is just one small part contributing to the cost of delivery of a service. A good way to state the difference between the services is - 1. A T1 provides a guaranteed "Full Commercial" level of service, 2. An ADSL connection provides only a shared "Consumer Grade" level of service.
So the reason for the price difference is all based on the cost (to the supplier) of the service required to maintain satisfactory operation.
Commercial T1s come with support and a short response time to fix. Also guaranteed bandwidth. If you're not getting the bandwidth you paid for, there will someone working on it until you do. And... due to the way the T1 technology works, in opposition to DSL circuitry, a T1 is faster due to less overhead. Not to mention that T1 guarantees T1 UP speeds as well as down.
It's not just the fact that it's a dedicated line, but the fact that it's a dedicated link running right from your premise to the demarc point of your provider. Also, business-class service comes with quality-service guarantees whereas consumer never has.
I don't know how they do it in other countries, but in the former SBC part of the U.S., The "T1" lines I have seen in the last few years are implemented as two bonded 786Kb SDSL lines. On the user premises, there is a converter to interface the DSL lines to the "standard" 4-wire T1 protocol, which goes to the user's CSU/DSU. The DSL lines go into a DSLAM and thence to the ATM cloud like any other DSL. Yes, there are still the service/bandwidth guarantees to justify the premium price, but it's not a dedicated four copper wires running from point to point.
Please clarify; as a general statement I have some concerns that this situation for ADSL is neither universal or prevalent.
In our region, Qwest's ADSL DSLAMs backhaul the data on fiber over ATM to the local ISPs; nothing is 'shared' or statistical about the muxing in this path; ADSL users get a PVC but still need to authenticate with PPPoA. The consumer actually gets the data rate promised and can get it as symmetric.
There is no general SDSL offering here but one can do SDSL over dry pairs point-to-point. From my limited reading of the literature, ATM PVCs on DSL beat T1 protocols for overhead efficiency.
Yes, you are correct in that no 2 providers would necessarily operate their DSLAM backhaul in the same way, so I guess it comes down to what type of service are they trying to provide from their DSLAM. However the reality is that a DSLAM aggregates DSL channels into large "pipes", and its the ratio of channels to "pipe" size that determines the ability of the service to provide performance comparable with a true T1.
However the aggregation of multiple DSL services to a single ATM feed is another form of Multiplexing, just not statistically. It all comes down to how they are dividing up the PVC's over the ATM, are they over-committing their bandwidth or not. If not, are their charges realising the cost for the operation of the environment, and if they are over-committed, then a single PVC has no guaranteed bandwidth. Of course most good providers will work out how much SDSL they can offer alongside ADSL and still maintain the levels of service they commit to, however one MUST acknowledge that while the bandwidth between the end user and the DSLAM is dedicated, from the DSLAM onwards the bandwidth IS shared. The real service difference comes down to how well the provider manages that shared component. If he gets the figures right, he may be able to deliver a "superior" grade of service, but no way can he cost effectively deliver a "dedicated" grade of service at regular consumer grade pricing.
Actually they don't HAVE to use PPPoA they can use other methods of authentication, but yes, PPPoA is the most common way to deliver ADSL. Of course SDSL is really low level and not even IP or Ethernet based.
The Bandwidth can only be "guaranteed" between the controlled end points (the customer premise and the DSLAM), because everything past the DSLAM really is a "shared" environment (EG the ATM). Then it comes down to how that shared environment is divided up (IE no. of PVC's per ATM bearer).
Yes, individually (IE per PVC) they can, but no provider is going to fully populate a DSLAM with SDSL service AND then connect an aggregate SDSL bandwidth with a backhaul that matches the DSL bandwidth, and sell that at non commercial (IE at consumer) rates. That would be shear lunacy, as all DSLAM manufactures provide a "recommended" mix of SDSL and ADSL in the one box based on cabling issues for each environment.
Even then, there is still a "shared" component, none of it is truly "dedicated" the same as a T1 is. I am not sure how you quantify the ATM overhead as "efficiency", its superior capacity is only there because it has surplus capacity available that allows it to deliver the required bandwidth. EG to deliver a 2.048Mb SDSL service the ATM signalling rate on the line is more like 2.6Mb. The resultant service can be near identical (and in some ways superior) to a conventional
2Mb service.
I prefer to think of it this way. With a "true" T1, the data is delivered over a single fixed "raw" pipe. With ADSL, a larger "massaged" pipe is used that has headroom to ensure the resultant delivered to the end user is the equivalent (volume wise) of a T1. Consider that to deliver T1 data speeds, an SDSL environment has to be configured at about T1 + 13% to ensure you really do get get the equivalent of T1 "end user" capacity.
Yes, it is possible to configure an "apparently identical" SDSL service that would out perform a "standard" T1 service, but then it is not possible to do this across an entire DSLAM platform and charge just DSL rates for the lot, the numbers just don't add up.
Locally (NZ) we use 2Mb (an E1) as the "standard" fixed pipe service. This can be delivered as a conventional 2Mb service, or via SDSL. The charge for either service is identical, and the performance is fairly similar, but as a large consumer of Leased Data Services here we know that there can be more "performance" related issues with the SDSL delivered data, particularly when the DSLAMS need to be restarted (after all they are more S/W based than H/W).
Another difference is that with DSL, there are usually *no* claims on latency. 200 msec to the gateway? As long as you can download anywhere near the rated speed, they probably won't even take a look at your line if you complain of high latency. With a T1, high latency will be a problem they will fix.
Another difference that probably is a big contributor to the cost difference is that T1 usually has an SLA (service level agreement)--basically, the contract between you and the phone company includes promises on their part to keep the line up and performaing to spec a certain percentage of the time, and to respond to problems within a certain time. There will be penalties if they fail to meet these requirements.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.