re: cache flows, dynamic versus static loads
LJM, my use of the name Akamai was a random pick, actually, and only representative of the group. I should have referred to the class of offerers - in this case being F5, AKAM, Sandpiper, and a couple of others - as a group instead of singling out any one of them.
In any event, I tend to agree with your assessment that the dynamic (and close to real time) feeds will present more of a problem for consistent delivery than those which are static and only change minimally over time. I referenced this fact, comparing a Britanica type of application to those which were streaming-based, in one of my prior posts.
This entire issue of cache channeling versus DWDM proliferation [did someone say bandwidth glut?], and cheaper bandwidth down the road, is one that intrigues me. On the one hand the netheads are talking about making the network stupid, and removing all intelligence except for the edge and the end points. And here, we see an elaborate set of schemes being implemented by the ASPs to use intelligence throughout, like it was going out of style. Perhaps the negative aspects of intelligence are only aimed at those previously associated with PSTN voice applications? I put my tongue in my cheek here.
In any event, and as you can see, I am not the one to be giving advice on a pick at this point in this space (which I ordinarily wouldn't do anyway), much less one that could be astronomical in price, when I myself am sitting on the fence, wondering what offsets might lay in store down the road. Hope the rest of the post helped, tho.
Regards, Frank Coluccio |
