Ahhaha, and Thread,
Ahhaha, you note that:
>>This is a primitive model. The WDM is the best Lucy and others can cook so far after years of big money development. The management, synchronization, and chromatic aberration go up geometrically with the number of colors needed to achieve the mixed above WDM throughput. <<
We should list the constraints that prevail today which are responsible for some of the limitations in existing systems, and discuss the modalities associated with various generic traffic types, if we are to make any sense out of all of this - to try to establish a backdrop for where this tech is appropriate, and when it might be appropriate in the future.
At some point common carrier (IXCs, xLECs etc.) will truly converge with cablecos at the organic networking level, possibly fostered by fat pipes and very fast IP flows, but this is not happening right now.
I therefore would like to handle this topic on the basis of how it currently might fit into the carrier model, and later at some point introduce ways in which the convergence, which has been elusive until now, might be facilitated by it.
SR's proposed type of solution is a point solution at this time, and one that is very limited given the universe of extant conditions that prevail in today's networks, from the physical, and all the way up the ladder to the application and management stages.
These conditions dictate acceptability by virtue of their need for interoperability and feature compatibility, and by virtue of the scope of various classes of users' applications and platforms that are capable of accepting such large flows, much greater, to say the least, than those which exist today.
Nevertheless, the SR model is one that needs to be explored, and exploited where possible, in those niches that it can populate, even if they are few and far between at this time, for this will foster assimilation into a greater mix of networked topologies with time. And this is not unlike the dilemma that WDM posed to carriers not long ago. They were simply unprepared to deal with wavelengths as recently as September of '96.
Let's put POTS, N-ISDN & B-ISDN, WLL and DSLs aside for the moment, since these are only end-station types of access solutions compared to what we're talking about here.
In the optical networking domain, one which to this point in time has been dominated by SONET, and before that the North American Digital Hierarchy of T-1s and T-3s, we have a number of speed-limiting factors that have relegated the top end of the line to 10 Gb/s aka the OC-192 rate in SONET terms. The more prevalent rate, however, is OC-48 or 2.5 Gb/s.
Most of these limitations are electrical in nature, and those which are optical are only now becoming pertinent or heightened at the awareness level, because the electrical rate problems are now slowly being solved. Very slowly, I might add. Some feel with proper justification that further electrical/electronic solutions will reach a point of diminishing returns, and eventually reach a point with no overall gains achievable, and then followed by loss.
See my preceding message to Silicon Investor for more on this.
Attempting to "fit" an SR type of solution to existing problems is difficult, because heretofore there has been a barrier in thinking as to what was achievable using such a model. Another reason is that the highway benchmarks have already been established, and the bed on SR's truck is too wide to fit into a single lane, it's been argued.
And therefore, the very types of problems that could be satisfied by such a solution were not articulated, because no one thought that such a solution was achievable or appropriate, where virtual pipe-sizing is an issue, in the first place.
This, I believe, is the essence of what goes beyond pure innovation, when a solution precedes an unknown problem, because the problem can only manifest after the solution is in place. ISDN followed, in a way, such a pattern, but unfortunately never lived up to the expectations of its promoters. [Should I qualify this with a 'yet'? Nah..]
There are other parallels taking place right now in the Internet Space that fit this type of syndrome. VoIP is one of them.
More interestingly, getting back to ISDN for a moment, it ultimately satisfied only a limited set of needs that were not even considered or known about at the time of its unveiling. And most of those center on Internet access, a phenomenon that took ten years to appear beyond ISDN's inauguration.
[[There are those who will disagree with my analysis of ISDN, typically die-hard ISDNers who have been near-institutionalized about its benefits, without the benefits of experiencing alternative means of access due to the BOX effective lock-out of such alternative availabilities to users, but the percentages and uptake rates speak for themselves: Dismal.
There are, in all fairness, a number of applications that ISDN satisfies very well, but these are by far inadequate when it comes to justifying the expense to cover R and D and the marketing efforts that went into their promotion. And all the while, who was paying for this folly? Rate Payers. And the loss statements? They showed up as Business As Usual, factored into the cost of doing business for the most part, with specified ROI revenues assured, unabated, for the most part. But I digress...]]
So, what are the other constraints? I'll let you (ahhaha) speak to those which are related to the quantum physics side of this problem, since without proper preparation, these analyses serve only to boggle my mind. Be gentle, though... will ya?
I'll focus, instead, on what I tend to gravitate to most often, and that is the mixture and texture of networks as they currently exist, and those which lie just beyond the horizon, and how such innovations as DWDM and SR's diffractive and compactive techniques portend to affect these environments.
The SR approach or one similar to Palmer's diffraction and compactification techniques obviously must be phased in over time, if it works as advertised, because not every topology [if any at all, at this time] currently calls for it, or, more to the point, if any situation could, indeed, "accommodate" it.
If we break up the telecom landscape into its dominant topological regions and application types as a function of reach, we can focus presently at the core, the edge, and the distribution & access components.
- the core is the densest part of the net at its center, handling massive flows;
- the edge is used to connect the distribution part to the core; and,
- the distribution part (commonly regarded as the outside plant part along with associated central office termination equipment, i.e., access technologies) handles user connectivity, aka, access.
And then there are further breakdowns at the premises, including enterprise LANs, and personal area nets, PANs (the latter, when combined with the usual form of POTS transport, collectively referred to as POTS and PANS).
If we eliminate Cable TV for the moment (which would require its own analysis due to still radically different metrics used for its delivery) as a consideration, and focus solely on today's dominant carriers' and corporate data users (enterprise) needs, I see the relative merit (i.e., penetration opportunity) of an SR type of diffraction and external modulation approach (giving them the benefit of the doubt with regard to their claims and explanations), superimposed onto a DWDM fabric, as follows:
Class........................1 yr..................2yrs......5yrs
Core.........................5%...................25%......90%
Edge........................2%...................15%......50%
Distribtuion...............1%...................10%......25%
These are obviously only my seat of the pants estimates, and they lag published WDM stats by some degree, but they should reveal my thinking of the potential that lies behind a unified beam approach wrt to acceptance and uptake rates. Other factors which could ultimately modify these uptake rates could conceivably result from displacement factors, whereby earlier "generations" of WDM might be prematurely retired in order to make way for the larger flows.
To review, what we're talking about here in simple terms is the potential utility of ultra-high capacity flows contained within singular streams, possibly multiplexed together on WDMs, as opposed to the convergence of many smaller capacity streams. Looking for comments and opinions here, folks.
|
