Idea: Microcell/PicocellFemtocell with Janet Backhaul
For the last three years, we have suffered from poor mobile telephone coverage in modern buildings on our campus which are constructed from materials providing better insulation (like foil backed Kingspan) in order to achieve BREEAM excellent certification.
There are a number of options, however they are not practical. For example you can get mobile repeaters; however there are licensing issues and you can get Microcell/PicocellFemtocells from specific providers; however they will accepted limited numbers of registered handsets and are provider specific.
It would be really helpful if Janet could investigate the possibility of partnering with a provider who could provide these devices (carrier agnostic) which we could backhaul over the Janet network to a peering/aggregation point and out to the respective carriers.
I know we are not the only people in the community with this problem, therefore this is an ideal opportunity for Janet to help solve a community problem and demonstrate partnership with industry.
Matt
Comments
I see we haven’t come back to you on this yet, so I thought I’d let you know what’s been happening around this idea. Your comments tie in closely with discussions we’ve been having within Janet for some time, and you aren’t the only person making suggestions in this area. However, there are some significant obstacles for us to overcome to deliver the functionality that might solve your problem.
Firstly, there are potential impacts to the Janet network. Short of requiring pre-registration of handsets that can use them, these H(e)NBs would be open to the public, and therefore the traffic they generate may risk compromising Janet’s status as a private network. We would also need to explore their status with respect to emergency service calls: the organisations deploying them and Janet could incur responsibility to provide accurate location data for the call point, and to guarantee high availability such as battery-backed power and fall-back options.
The mobile provider’s networks are also a factor. Despite mitigation techniques, H(e)NBs typically can’t be deployed where the outdoors signal strength is reasonable, because of the risk of co-channel interference (even if as in the example you provide, indoor materials limit signal strength). This is another reason that the location of the devices must be registered and they cannot subsequently be moved. H(e)NBs operate in licensed spectrum, and their use is tightly regulated.
The solution you suggest requires collaboration between competing providers. It’s possible that we could act as the catalyst to establish this (although the Janet community isn’t quite big enough a market to compel them), but as far as we can tell it hasn’t happened in the global market yet – indeed the deployment of femtocell technology overall is far below the levels that were projected for it back in 2007/8. Fundamentally, under the suggested model one of the partners would have to donate licensed spectrum that they purchased at significant cost in the 3G auction in order to facilitate their competitors extending footprint and gaining additional traffic income. The device would also have to support multiple configurations for the various non-standardised methods for passing traffic back to the various suppliers’ core networks.
Finally, there are strategic considerations here. In terms of data mobility, at sites with good backhaul we would certainly recommend eduroam Wi-Fi as a higher bandwidth, higher quality, free-at-point-of-use option over facilitating expensive 3G/4G data – and the providers favour offloading data to someone else. We also predict, with the spectrum auction now set for this autumn, that the rollout of 4G/LTE, especially in the ‘whitespace’ bands, will substantially reduce the number of ‘problem’ buildings, as the lower frequency signals have greater penetration. So, we could be building a solution to a problem that changes greatly within a 1-2 year timeframe.
So, overall I’d say we see the value to helping you solve this problem, but we’re wrestling with a few aspects of this, and a compelling solution hasn’t come to light yet. All input to help us with navigating the issue would be greatly appreciated, of course.