LTE Heterogeneous Networks & Carrier WiFi: What is the opportunity?
View the Q&A
WiFi is becoming a key concern for network engineers as more and more wireless carriers are considering it as a cost effective solution for tackling skyrocketing consumer data demand (and cable operators wishing to expand their fixed broadband market appeal). While attractive for its simplicity and wide adoption in user equipment today, WiFi, nevertheless, is no silver bullet in terms of the fundamentals of telecommunications. Like other systems, it is not interference-immune and its capacity has limits. The evolution of the standard offers promises, but user equipment is years from being able to leverage that evolution. This webinar will address the design challenges of a carrier WiFi network and assess its potential as a data offloading solution for existing 3G networks.
Join us for this free webinar as we address the following:
- Identifying where and how to deploy WiFi (indoor, outdoor)
- Designing a carrier WiFi network using outdoor hotspots in a metropolitan area
- Assessing the potential for data offloading from an existing 3G network
Tags; Archive, Carrier WiFi, Heterogenous Networks, LTE, LTE, Mentum, Networks, telecoms.com
Thank you for all your questions today. This Q&A is now closed. Please send your questions direct to Bernard at bernard.breton@mentum.com.
You mentioned the integration of Wi-Fi to a carrier’s macro network will be improving – what are your thoughts on what needs to be in place so that integrated Wi-Fi can become a truly manageable asset? Can it be network based only or will it require a combination of new client capabilities in the device + network? Very good presentation!
This is a great question. I think both approaches could work. Some vendors right now are offering device based approaches. The issue with this is that there is a lot of devices out there (old, current and new) and a device-based approach is hard to implement. Hence, a network approach, while more complex, would certainly be preferable. I suspect we will see both approaches for a while but that ultimately, the network-based approach will dominate, just like it does in the cellular networks.
what are your thoughts on combined public hspa femto/wi-fi cells vs. if i understood correctly pure wi-fi?
In this case study, it was a pure WiFi offload but a cell combining several technologies such as HSPA/LTE/WiFi would offer a great level of flexibility. In this case here, these were really pico cells rather than femtos.
What assumptions did you use for the wi-fi backhaul capacity? I believe this impacts the number of wi-fi cells you need for offloading.
The assumption I used was simple: I assumed unlimited backhaul capacity as I focused on the air-interface capacity. That being said, if you have limits on the backhaul (which limits depend substantially on the backhaul strategy itself), then this can certainly increase the number of WiFi cells required as this would basically mean that the WiFi cell capacity will be below their air-interface capacity.
could you provide some background on the type of wi-fi cells deployed and how they were or were not integrated to the carrier network?
In this case study, we used primarily lamp-post installed WiFi cells with 2 antennas per site (basically one on each side of the lamp post). There were about 20 cells which were side mounted on buildings. They were all considered available for offloading and they were not carrier traffic other than the offloaded traffic. In terms of handover, we assumed that all user device that was covered by WiFi was using the WiFi network as a priority unless the WiFi cell was overloaded.
How good is immunity to interference from other nearby networks and other sources
The question of immunity to interference is really dependent on which “flavor” of WiFi we are talking about. The IEEE is working on improving this as WiFi is very sensitive to interference. The 802.11v is an amendment which should enable better coordination of interference through better knowledge of the RF environment.
Do country regulators have an important role to ensure licenced operators have been given spectrum that is useable.
Regulators are providing usable spectrum to operators when it comes to licensed spectrum. WiFi is using unlicensed spectrum. This ultimately, is one of the major difference between deploying WiFi and LTE small cells.
How do I get a copy of the presentation for my records???
The presentation will be posted on the Mentum web site in the coming days at http://www.mentum.com.
Do you see an increase (%) in total traffic handled due to Wifi deployment?
Great question. I did not make that assumption in the analysis but from several studies and research, it is certainly suggested that right now, users do tend to use much more data when connected to WiFi. In my view though, this is primary related to the impression that when they are on WiFi, it’s free. In a model where you offer data connectivity experience and price which is the same across the technologies, I suspect this difference would disappear. In any case, the analysis could have been made with a different set of assumptions.
Has any thought been given to simulcasing microcells or nanocells to improve spectrum efficiency?
No, I did not really gave any thoughts to such approach.
What propagation models are used to predict and simulate in 3D models
In this example, we used the Mentum Universal Model, which is the best model Mentum has in its portfolio. It’s worth noting that we used a mix of 3D and 2.5D in this example.
I understand large number of wifi cells will require an extensive FTTX network. Will this result in making the wifi solution more expensive than reported in your presentation?.
I did not really look at the cost of the WiFi nodes in this presentation but the cost of the core/backhaul is a key consideration when the number of cells explode, irrespectively of the technology selected. I believe that NLOS backhaul solutions might prove the most financially realistic, particularly when the equipment is co-located with the macro-sites.
Thanks, Bernard !
Was interference with non-operator wifi access points considered in this simulation?
I assumed an average 3 dB degradation on the downlink. Indeed, the challenge with WiFi is that you do not know where the foreign nodes are and what their loading is. Hence, you have to rely on either statistical approaches or a link budget approach (which is I did).
As number of cells go up, operator need to take less responsibility for site (like power, rent etc)..femto can be a good opition…but in case of wifi we need to take such responsibility
what do u think?
In my view, the primary challenge with femto cells (here I really assume cells that are installed by the end-user in their premise) is that it is decreasing the perceived value of the service by the end-user. As long as the decreased in perceived value is matched by an greater decrease in cost, then this is fine from a pure financial perspective. Ultimately though, I believe that what the mobile operator truly offers is ubiquitous data connectivity to devices. If you start to involve the end-user in providing their own (at home or at office coverage), I am of the view that is ultimately undermining the core business of the operator.
Are the WiFi cells usually restricted to the customers of the operator that provided them and what if any charging mechanisms are used?
The assumption in this case study is that the WiFi cells were not carrying “foreign” traffic. It would certainly be possible to work on the basis of other assumptions. Given the capacity of WiFi and very short cell radius (50 – 200m), this should not change the analysis too much. Another option would be for a WiFi operator to offer offloading to several carriers. This would certainly be possible given WiFi capacity but it would require a quite careful planning of the WiFi cells.
How do you envision managing spectum concerns (i.e. interference, reuse, etc.) in the model you presented. I understand that the current WiFi ffering is unlicensed, but I still do not understand how you intended to manage spectrum of the light pole sites that you presented. Is or was there a specidic channel plan employed?
I considered a 3 dB average downlink degradation due to foreign interference. The channel plan for the WiFi nodes that were deployed was managed to control interference. This cannot be done with foreign WiFi cells.
How was the wifi coverage modeled and tested ? what was the typical height of a wifi access point ?
The assumption was lamp-post installation. The modeling was done using Universal Model along with a mix of Cityscapes HD geodata, which is a height-attributed clutter and full 3D geodata (for the core).
Hi Bernard, thank you for a very stimulating presentation.
I am very familiar with the Dallas CBD – it is somewhat unusual, as it is virtually deserted between 6:00pm in the evening and 8:00am in the morning. I assume that your capacity analysis relates to the business day period only and that the economics of provisioning Wi-Fi offload capabilities are attractive for relatively concentrated (by time of day) network demand?
The traffic assumption was based on the peak load during the day, which in the Dallas core would be during the working hours. The traffic density is therefore highly linked to the total office space.
When do you expect seamless cellular/Wi-Fi handover to be available?
The key to this is the 802.11-2012 cumulative standard which was recently ratified. This includes 802.11ac but no equipment yet support it as far as I am aware. In the past, vendors have been pretty quick at supporting new releases to 802.11.
Will wifi ac handoff? Is it that different? You can buy the routers now as of a few weeks ago.
So, the cumulative standard recently ratified includes handoff capabilities?
Yes, it does support the 802.11u amendment.
802.11ac offers no specific support for handoff. It is really 802.11u which is an amendment to the IEEE 802.11-2007 standard to add features that improve interworking with external networks. This is part of the 802.11-2012 which includes 10 amendments to the 802.11-2007.
Can this type of study be extended to other outdoor public access small cells? (other than wifi ?
Absolutely, we could have done a similar study using LTE small cells rather than WiFi.
What are your thoughts about the future of LTE small cells?
This case study was about WiFi but clearly, LTE small cells is a very solid option for operators. The analysis done would have been similar. The primary difference is shared spectrum (WiFi) vs. licensed spectrum. The 2nd most important element is that LTE is truly a carrier-grade technology.
How was wifi backhauled ? and was there any backhaul limitation (throttling)…also were all wifi cells omni directional antenna pattern or did you experiment with sectorized pattern
I assumed no throttling on the WiFi nodes due to the backhaul. All pattners where 90deg. with 2 sectors per lamp-post.
why femto unmanaged?
I considered Femtos as un-managed as they are often end-user installed.
Did you consider the impact of indoor WiFi hotspot? Is that something we could model in Planet?
I did not consider that this area had indoor WiFi hotspots but this could have been modeled quite well in Mentum Planet.
Which Mentum Planet release do I need to do WiFi modelling?
You need Mentum Planet 5.5.