Aricent recently hosted a webinar titled Next Generation Backhaul for Small Cells. The webinar provided Telecom Equipment Manufacturers with exhaustive insights into small cell backhaul requirements and a roadmap for how backhaul networking requirements will evolve to meet the unique demands of the public access small cell era. We discussed the architectures and features of various public access small cell backhaul technologies and identified the key considerations to arrive at the most optimal mix for different small cell backhaul requirements.
Given the large number of questions asked we were not able to answer every one of them live, so we've addressed all of them below.
Is Open Access Femto considered to be a Small Cell technology by your definition?
Yes, it is. Small cells encompass femtocells, microcells, and picocells. The focus of the webinar has been on the public access cells (which are open access cells indeed).
Security also should have been a consideration for BH selection, particularly for public BH?
The current cell towers have the same amount of exposure as the small cells with respect to security. With regard to data security, the access between the user and the cell is much more vulnerable than the backhaul network itself. Since all the schemes have the same level of exposure with respect to security, we did not specifically include this.
What is expected to be the geographical span of a typical single cluster of Small Cells i.e. the group of small cells that are connected to the same macro site?
In urban areas, which are expected to see the maximum usage of small cells, this should be within the half-mile range.
What is the capacity required for the small cell on the different topologies and at the aggregation points?
Each small cell will typically be expected to handle the traffic from approximately 25 users. When you assume smart phone and data prevalence, and use an over-subscription ratio of 1:3, this translates to slightly less than 100 Mbps per small cell.
What will be the impact on synchronization in the backhaul network to meet the demands of small cell network deployments?
I assume you mean “time synchronization”. This becomes absolutely necessary and stresses the need for technology such as synchE and IEEE 1588 all the more.
What is the general consensus on the need for demarcation/OAM/fault finding of such small cell networks?
These are absolutely necessary since the service provider cannot compromise much on the quality of service or quality of experience to the users. Lack of OAM techniques results in an unpredictable network and the cost to maintain would become much higher for the operator.
Would operators generally deploy L2 or L3 demarc?
We expect operators to deploy L2 demarc considering the simplicity, ease, and cost, coupled with its performance benefits.
Power consumption has not been analyzed in your comparison of wireless/wired backhaul architecture. Is it relevant for the choice of the backhaul solution?
This is a good point. It definitely is relevant. At the same time, other operational costs and maintenance costs tend to outweigh the power consideration in this aspect.
Any concerns with latency?
Wired media in general lend themselves to lower latency, so, this is a concern with wireless media. Use of microwave for backhaul has, however, been quite successful and allays this concern.
Where has Aricent rolled out this service?
Aricent does not roll out services to end users. Aricent provides frameworks and professional engineering to equipment manufacturers and operators to realize the service. Aricent is working extensively with customers in North America and Asia in this area.
What resilience is required i.e. is battery back-up for public small cells required?
These are more likely to be powered from the power line itself. In geographies where power outages are common, battery backups will be needed.
Do you see the NLOS MW as a solution for the backhaul? How good is it?
Operators have been successful with this. In fact, the sub-6 GHz non-line-of-sight (NLOS) microwave link with a point-to-point multipoint architecture is an operator favorite.
What is your view about one box solution vs. two box solution for RAN and backhaul in small cell?
Operators look for more than the “less equipment” line of argument. The total cost of ownership and the utilization levels that the solution provides are more important to operators than just the number of boxes.
Why does Aricent state that Ethernet OAM is not important for small cell networks? Don't small cells connect backhaul via Ethernet? Aricent’s position is that the OAM is indeed a vital necessity. Why this is not stressed is because the techniques have become reasonably well-established at this time.
Is it true that fiber may be preferred to microwave if the small cell is placed on the building wall, and there is fiber present in this building?
If there is existing cable laid, it is always preferable to use it instead of the wireless option. The wired option is inherently more reliable and provides more capacity.
What is your take on WiFi-based backhaul as a viable option? Are you aware of deployments by providers using it as a last mile option?
This has started seeing traction only recently. It required the development of adaptive directional antennas and smart mesh techniques to make this a viable alternative. There are some operator trials but no established deployments that we are aware of yet.
PmP microwave has not been mentioned. What role do you envision for this technology as far as small cell backhaul is concerned?
We did not specifically distinguish P2P microwave and PmP microwave. In fact, PmP microwave is a popular selection with operators.
I understand that xDSL standards' jitter specifications are unsuited for LTE backhaul. Is that true?
The jitter in xDSL lines varies in different geographies. At the tolerance levels of the specifications, yes, it is unsuitable, but not all DSL rollouts have high levels of jitter. When considering DSL as an option, an operator would have to evaluate the jitter aspect critically.
Is synchronization (Phase/Time) an immediate requirement for small cells?
Yes. This is required especially for wireless technologies to make sure that the different nodes use the same frequencies and clocking, and it is imperative for the correct operation of the network.
From your experience what is the small cell vs. the macro cell? In another words how many small cells on an average will have to support a single macro cell?
The average is in the range of 10 to 15 small cells to support a macro cell.
What will cache-integrated small cells do for back haul and the other way around?
Caching improves the adoption of small cells for data dominant usage but increases the footprint and cost of the small cell.
Question for DAS, is it a Small Cell in your definition? Are they competing solutions or complementary?
It is not a small cell by our definition. Small cells and DAS are competing alternatives.
How would synchronization requirements be met in small cell networks? What would be the most popular choice GPS, PTP, Sync-E etc.?
GPS is expensive. We expect a combination of PTP and Sync-E to dominate.
What are the targets for delay? The last mile backhaul chart indicates that delay is high for cable. Testing has shown that Cable is capable of very low delay.
Cable being a shared medium has inherent unpredictability. So, during lab tests, results may be promising since the bandwidth in the cable is high, but when actually deployed, the results may be far from what is seen in the lab. This is the reason (non-predictability) that the delay is shown as high.
Which MEF standards standardize reliability and resiliency of the CE service?
MEF 2 describes protection requirement. MEF 10 and MEF 10.1 describe the service attributes. MEF 22 lays out the requirements for a backhaul network. These refer to standards ITU-T G.8031 and ITU-T G.8032 for the actual resiliency mechanisms.
Femto cell technology is well established. Since these provide the required benefits, why do operators need small cells at all?
The rollout of a non-open-access-femto is daunting and very expensive affair. The open access femto, however, comes under the definition of the public access small cells that we have covered in this webinar. The open access femto is indeed the “small cell” that we are discussing.
What are the ways to speed up the installation of a small cell? Can you outline a few?
Auto-provisioning and topographical planning for placement are 2 key factors that can significantly reduce installation time.
How much time does it typically take to install a small cell versus a macro cell site?
For a macro cell site, the time taken is in weeks but for a small cell, it is in days or hours.
What are the radiation hazards of planting so many antennas all over the place? Isn’t that prohibitive?
Since the power output and reach of each small cell is a fraction of a macro site, the radiation hazards are expected to be within tolerable limits. WiFi is, anyway, ubiquitous today, and the radiation from the use of WiFi has been concluded to be non-hazardous.
What about security issues? Both physical and data security? Since small cells are in public locations, aren’t they very insecure? Would not data going through small cells be compromised easily?
Data going over the air between a mobile device and a macro site has the same exposure as the data between the mobile device and a small cell. From this perspective, small cells do not create any additional data security concerns. Strong authentication mechanisms are in place currently between cell sites and mobile devices and the same is expected to be sufficient for small cells as well. Physical security is a much bigger issue and currently there are no established or proven ways to guarantee this in a uniform way across the world. So, you are right, this aspect is a key factor for small cell deployment in several geographies.
How is timing and frequency delivered to small cells? You don’t seem to talk about this aspect in your slides. Isn’t this very critical and important?
Yes, this is going to be critical, since the timing to user devices will be delivered from the small cell and also because for wireless backhaul, all the small cells must be synchronized to the same frequency. 2 key technologies that address this and are getting widely deployed are SynchE and IEEE 1588. We expect these to become defacto implementations in the small cell networks.
Will not frequency delivery/synchronization to small cells involve running a wire to them? So wireless technology to connect the small cell to the rest of the network cannot work? Or at least cannot be reliable?
IEEE 1588v2 is capable of delivering this requirement and can avoid running a wire to every small cell just for synchronization. Since the small cells are a single hop or a very small number of hops away from the macro site problems of larger networks, like large delay variations, do not occur with small cell connectivity and therefore IEEE 1588v2 becomes viable without SynchE in this environment.