MCUs & MPUs
but the CAPEX/OPEX expended could erode a significant part of the benefit gained from freeing up bandwidth and adding more users. Therefore, any enhancements should be made in such a way as to keep increased expense to a minimum while achieving the biggest return in terms of network usage efficiency.
Maintaining algorithmic
agility: ROI coding is an active research field; new techniques and algorithms are continually being developed. Therefore, there is a need to maintain a level of flexibility to adapt to these improvements and enable software upgrades throughout the CDN.
Figure 2 - before Figure 3 - after
Choice of compute platform When the above factors are evaluated, it becomes clear that the x86 architecture is not optimised for the large-scale signal-processing demands required by advanced video processing. If applications are miss-targeted to this platform, it can result in rapidly increasing data centre operational costs. To illustrate this, Figure 4 shows two scenarios: Processing the world’s video using x86 and processing video traffic using the LSI StarPro multicore media processing Digital Signal Processors (DSPs). Service providers can further manage their compute expense by linking to advanced edge caching schemes to prioritise which content would give the biggest spectral return when processed with advanced bandwidth-reduction schemes. The most obvious technique would be to prioritise based on number of accesses. For example, a popular sports event being viewed by multiple users will return much more resource benefit compared to obscure long-tail content being viewed by a single user. If service providers improve a standard 1.5Mbps stream being accessed by five users in a cell by just 10 percent, the return would be 750kbps. In a fully loaded cell, this would represent enough bandwidth reduction to allow around 10 – 20 web users or around 100 voice calls, a significant benefit to the service provider.
Figure 4
Ability to support a significant proportion of traffic: As video usage continues to grow, the number of streams grows, as does the bandwidth required. Applying the meta-codec techniques described must be done on as many channels as possible to gain maximum benefit. This may have to be done real time or it could be done non-real time and stored (in some cases), but either way it will have to be done at some stage. Managing service TCO: The result of these advanced techniques may achieve more efficient use of the air interface,
www.cieonline.co.uk
Mobile service providers are seeing the potential in mobile video to increase average revenue per user, but there is a need to manage finite physical resources, specifically, the radio spectrum. Mobile service providers will have to drive the adoption of more complex bandwidth-reduction schemes. DSPs can offer a compelling way to significantly reduce the computational CAPEX/OPEX required to implement these advanced schemes. By reducing this
expense and by maximising the efficiency with which CDNs use the RF bandwidth interface to the end user, mobile service providers can continue to add to the rich variety of services available for end users.
LSI Corporation |
www.lsi.com
Richard Benson is Senior Product Line Manager, Network Components Division, LSI
Components in Electronics April 2012 29
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64