SoylentNews is people
SoylentNews
from the well,-they-would-say-that,-wouldn't-they? dept.
'This breakthrough is very exciting for us. In just six months we have developed a truly unique demonstration which is able to use standard IP end-points and translate the IP flow into an IP-over-ICN abstraction (publish/subscribe model), where a single hop within the ICN network used SDN switches with pre-installed forwarding rules for the ICN flows,' says Dirk Trossen, Principal Scientist at InterDigital and POINT's Technical Manager. The POINT project kicked off in January 2015 and is supported by the EU to the tune of EUR 3.5 million.
ICN, which stands for Information Centric Network, has been the focus of many research teams since 2006 for its potential to replace the IP-based Internet as we know it. ICN notably allows for content to be fetched from multiple servers and caches, for savings in the synchronous and quasi-synchronous delivery of popular content, for securing the content rather than the endpoint, and for allowing operators to apply traffic engineering rules.
However the requirements of such a switch – including heavy standardisation, strong stakeholder support and the fact that viable methods to create a truly scalable internetworking of individual ICN highlands has not yet been found – have cast doubt on its feasibility.
To overcome this obstacle, POINT tries another approach. Instead of seeking to replace Internet Protocol (IP)-based networks, the project aims to harness the innovation potential of IP-based applications and solutions, while benefitting from specific ICN solutions in terms of their potential for better performance compared to their IP-based counterparts. At its core is the fundamental question: Is an IP-over-ICN system a better solution for IP-based services than pure IP-based networks?
A customer use case in the project's presentation flyer helps to illustrate the POINT approach: John, a London priest wanting to reach out to older parishioners whose health doesn't allow them to attend services in person, would like to set up a live video streaming service to fulfil this need. Unfortunately his bandwidth is not high enough, and paid streaming services are too expensive. He reluctantly chooses to use YouTube despite his fear of losing his content rights and sends over the data, thereby enabling a high number of users to view the video simultaneously. With the POINT software, John could have created a unicast stream received by users as a multicast stream, meaning he wouldn't have to worry about his bandwidth anymore, and could do without both paid and free streaming services.
(Score: 0) by Anonymous Coward on Saturday August 01 2015, @04:46PM
Back in the '90s, ATM (i.e. Asynchronous Transfer Mode, not cash machines) was supposedly destined to become the backbone protocol of the Internet, either displacing or carrying IP packets between boundary points of the carrier networks.
Whether it actually happened or not (I think not, but I'm not sure) turned out to be of negligible concern to 99 1/2 percent of all IT workers.
(Score: 1) by Hyperturtle on Saturday August 01 2015, @05:44PM
I think it sounds like something we hear every few years, just like how this version of Windows is truely the most secure version and best ever.
The message doesn't change when someone has an idea or product to sell.
SDN is just a GUI and scripting for configuring network hardware. Large places may find cost savings by firing the people managing the hardware and having a programmable robot do it. Some smart person will be on-call for when it breaks. That person will be expensive, but will do the work of a number of people. CCIEs, etc are the types to remain, or just demonstrably smart people. 5 years from now, SDN will be available with everything if not enabled and used on everything. It has nothing to do with replacing IP or anything like that, it's just a way to make it cheaper to install new hardware to help display ads, when it comes down to it. It's not like the layoffs reduce prices for the services.
ICN is caching and intelligent routing based on the source and destination using additional metrics to determine how to best route that traffic.
Often, between networks, the honor system for metrics (if any are even used among peers besides for link speeds) is pretty horrible. Inside an ISP or autonomous system, the ISP may actually choose to do a good or careful job of metric calculation. However, if they are accused of setting metrics in their favor (not just QoS), everyone screams if someone's interests are not met as a result.
It would make sense then to host CDN caching servers and then provide a result on how to reach them to remain inside the network, right? or video or audio or whatever to take the best path (which may not be the shortest, or the fastest, or through the least amount of hops -- this is where metrics come into play and not QoS [but QoS can influence the traffic before and after metric calculation and decision making])...
It's different than the internet 2 that was tossed about and never really arrived. I'd like an internet 2 to replace what we have now, but this is not it. The examples are capable on a network now were it not for capitalistic gain being the obstacle. Instead of already being able to do the things this problem proposes to solve (which all were already possible if the services were actually available from everyone's ISP) I (we) have an internet that depends on my info rather than providing me free academic research materials that may help further me in my career. Those are locked behind paywalls instead of being ad-driven. Surrendering my data doesn't even help make me a better person. It just tries to get me to spend money on empty consumerism in exchange for some facade of what the promises used to be like.