As far as I recall, 'femto' in terms of unit means '10^-15'. You would notice this represent very, very small number. From this, you can guess 'femtoCell' is 'very, very small cell', 'very, very small eNB'.
When you say 'eNB' you would imagine such a large towers with three antenna spaced at every 120 degree, but the femtoCell is very small size compared to the normal eNB (they usually call this 'Macro Cell').
How small it would be ? It would be just a little bit larger than your cable modem or WLAN Access point you would have at your home.
The motivation to femtoCell is to develop a small sized eNB that can be installed within a house like cable modem or WLAN Access point.
Four questions follow, and the page asks them in its own words below. What a femtoCell can do, why an operator would want one, how it reaches the rest of the network, and what it makes harder. 3GPP has since answered two of them.
- Does a femtoCell do everything an eNB does ?
- Why put a femtoCell in the network ?
- How does a femtoCell reach the rest of the network ?
- What does a femtoCell make harder ?
- Reference
Does a femtoCell do everything an eNB does ?
The bullets below answer yes, and the specification agrees with them. What 3GPP adds is a name and one limit, and both are worth having before the later sections use them.
Does femtoCell has same capability as normal eNB(Of course, RF power would be much small... I mean the capability except RF power) ?
- Yes... original intention is 'femtoCell has full eNB capability as normal one'.
- Transmission Power should be much lower than normal (Macro) eNB since it is designed to cover much smaller area.
- Reciever sensitivty and dynamic range may be smaller than eNB due to small coverage
- Number of subscribers (UE) that can support is smaller than eNB. (I think in most case a Femto Cell would support less than 10 UEs even though some specification shows it can cover up to 20 or more)
3GPP does not use the word femtoCell. In 36.300 the device is a Home eNB, written HeNB, and a cell it serves is a CSG cell when access to it is restricted. The industry term and the specification term point at the same box, and a reader searching the specifications needs the second one.
On capability 36.300 is blunt. A HeNB hosts the same functions as an eNB, and the clause then lists additions rather than subtractions. The bullets above are right that the difference sits in the radio and in the scale, not in the protocol stack.
One limit belongs beside that list. 36.300 states that one HeNB serves only one cell. A macro eNB routinely runs three sectors or more, so a femtoCell is not simply a smaller copy of the large one.
A femtoCell is a Home eNB in the specifications : 36.300 uses HeNB, and CSG cell for the restricted case, which is what to search for.The protocol stack is the same one : 36.300 gives a HeNB the same functions as an eNB and then adds to them.One HeNB carries one cell : the box is limited in a way a three-sector macro site is not.The difference is radio, not software : power, sensitivity and the number of users served are what shrink.
Why put a femtoCell in the network ?
The two reasons below look separate and are one argument seen from each end. A short radio link needs less power for the same received quality, and both the network and the handset collect part of that saving.
Why they want to introduce this femtoCell into the network ? We can easily think of several motivation
- i) in High data mobile network, usually a cell radius is very small and to guarantee the signal quality at cell boundary is crucial to achieve such high data rate communication. To achieve this kind of condition, they need huge cost and time to determined the location and install the normal cell tower. But if they can put a lot of small scale eNB at home like WLAN Access point, they can secure the signal quality with very low cost.
- ii) If we can get the high quality signal with femtoCell even at shadowy/cell boundary area, they can reduce the transmission power of the normal eNB which in turn can save a lot of mony in terms of engergy consumption of system operator. (Of course, you should pay some of this electricity charge in stead of system operator if you have a femtoCell at your home).
Path loss is what drives it. Loss rises steeply with distance, so a femtoCell a few metres from the UE reaches it with far less power than a macro cell a kilometre away. Nothing about the modulation or the protocol changes; only the distance does.
That is why the case is usually made twice. The operator gains usable signal where a macro cell struggles indoors, and the handset gains battery life because its own transmit power falls. The bullet marked ii above takes the operator's half of that and counts it as an electricity bill.
Distance is the whole of the argument : a metres-long link and a kilometre-long link need very different power for the same received quality.The saving appears at both ends : the network spends less power reaching indoors, and the UE spends less reaching back.Indoor coverage is the hard case a macro cell cannot fix : walls attenuate, and more macro power is an expensive way to answer that.Nothing in the air interface changes : a femtoCell gains from geometry rather than from any new radio technique.
How does a femtoCell reach the rest of the network ?
This is the question the page leaves open, and it no longer has to be. The bullet below dates from the releases where the work was still in progress. The architecture has been specified since, and Figure 1 draws it.
How a Femto cell can be connected to other network components (like other Macro eNB or MME and other Femto cell) ?
- Up to Rel 8/9, there seems to be no clear specification of these connectivity, but in Rel 10/11 3GPP start defining these connection interface and finally it seems that a Femto cell is using almost same 'connection interface' as a Macro eNB.
36.300 clause 4.6.1 gives the shape. An operator may deploy a Home eNB Gateway, and that gateway concentrates the control plane. It exists so that the S1 interface can carry a large number of HeNBs without every one of them terminating separately at an MME.
The useful part is that the gateway is invisible to both sides. 36.300 puts it in two sentences. The HeNB GW appears to the MME as an eNB, and it appears to the HeNB as an MME. The S1 interface is the same whether the gateway is there or not, which is why a HeNB needs no special S1 variant.
The user plane does not have to follow the same path. S1-U may terminate at the gateway, or the HeNB may hold a direct connection to the Serving GW.
Figure 1. How a femtoCell attaches, from 36.300 clause 4.6.1. The gateway is a control plane concentrator that neither side can see, the user plane may go straight to the Serving GW, and X2 runs directly between two HeNBs.
- The two solid arrows are one S1-MME path in two hops. Neither end knows the gateway is in the middle.
- The dashed arrow is S1-U going straight to the S-GW, which 36.300 allows as an alternative to terminating it at the gateway.
- The vertical line is X2 between the two HeNBs, and it does not pass through the gateway at all.
- A HeNB connects to one gateway at a time. 36.300 rules out S1 Flex here, so a HeNB never holds connections to two gateways or to a gateway and an MME at once.
The second half of the question has an answer too. 36.300 states that this version of the specification supports X2 connectivity between HeNBs, whether or not either of them sits behind a gateway. One femtoCell can therefore hand a UE to a neighbouring femtoCell directly.
Which of those handovers is permitted depends on who is allowed into the target cell, and the next section comes to that.
The gateway is a concentrator, not a new node type : it looks like an eNB to the MME and like an MME to the HeNB, so S1 is unchanged on both sides.Control and user plane may take different paths : S1-U can bypass the gateway and run straight to the Serving GW.X2 between femtoCells is specified : 36.300 supports it regardless of whether a gateway is in use.One HeNB, one gateway : S1 Flex is excluded, so there is no second path in reserve.
What does a femtoCell make harder ?
Four concerns are listed below, and two of them now have specified answers. Neither answer is the one the bullet guesses at, which is what makes them worth reading side by side.
Isn't there be anything to be concerned ? or anything that may cause any side effect ?
- i) Most cricial issue would be "Who would install those femtoCell if I want to have it in my house ?". Would the network operator send an engineer to my home to install this ? or I have to install it myself ? If I have to install it myself ? How can I configure it ? I am not a trained people to configure this kind of thing. Just plug in the power would be OK ? I don't think the network operator would send you the people to install it since it would be costly. You might have to install it by yourself. Then what about the configuration of the cell ? That is why they are talking a lot of SON(Self Organizing Network). Basic idea is that once you just plug in the power, the femtoCell automatically configure itself to provide service.
- ii) If femtoCell works like a normal eNB, it should have its own Physicall Cell ID so we may need a huge pool of Cell ID, but LTE cellID is only 504. How we can handle this kind of cell ID saturation ?
- iii) Usually a Femto cell is owned privately. So the owner would not want to allow everybody to get access to the cells. He/She would may want to set of specific users to get access to the cell. How to handle this kind of situation. The concept of CSG would handle this issue.
- If you are arount a Femto cell but you are not allowed to get access to it. In this case, the signal from the Femto cell can be a strong interference source to you and you will be a serious interference source to the Femto cell. This would be the most critical issues which everybody is looking for solutions to. ICIC/eICIC can be a strong candiate for one of the solutions.
- iv) On the aspect of mobile phone maker, would this femtoCell introduction can bring in any technical complication ? Yes, it is highly possible and we don't know now exactly what kind of complication it would introduce.
The second concern asks for a larger pool of cell identities. 3GPP did not enlarge the pool. It partitioned it. In 36.300, CSG cells on mixed carriers broadcast a range of PCI values that the network has reserved for CSG use. A UE can then tell which identities might belong to a home cell without testing all of them.
That reservation is advice rather than a rule. A UE treats the last received range as valid for at most 24 hours across the PLMN, and 36.300 adds that what a UE does with the information is implementation dependent.
The third concern asks who is allowed in. The answer is the Closed Subscriber Group, and 36.300 names three access modes for a HeNB: closed, hybrid and open.
The hybrid mode is the interesting middle. 36.300 defines a hybrid cell as one broadcasting a CSG indicator set to false together with a specific CSG identity. Members of the group reach it as a CSG cell, and everyone else reaches it as an ordinary cell.
Access mode then decides mobility. The X2 handover table in 36.300 lets any eNB or HeNB hand a UE into an open or hybrid access HeNB. A handover into a closed access HeNB applies only for the same CSG ID and PLMN, and only when the UE is a member of that group.
The identity pool was split, not grown : a reserved PCI range is broadcast so a UE knows which identities may be CSG cells.That split is a hint with an expiry : it holds for at most 24 hours, and acting on it is left to the implementation.Access is a group, not a password : the Closed Subscriber Group decides membership, in closed, hybrid or open mode.Mobility follows access mode : a handover into a closed cell needs the same CSG ID and PLMN and a UE that belongs to the group.
Reference
For some further details in terms of business model and some technical issues, refer to following links.
- LTE Femto Intro, Femto Forum
- Driving & Drivers of the Femto Forum
- Energy-Efficient Mobility Management for the Integrated Macrocell-Femtocell LTE Network
- Recent Trends in Femtocell Research
- Femto LTE-A Hardware Design Challenges (Very Good Hardware Overview)
- Interference Management in Co-Channel Femtocell Deployment (Very Good Overview and Details)
[3GPP] 36.300 : 3GPP - E-UTRA and E-UTRAN; Overall description; Stage 2, v19.2.0. Clause 4.6.1 gives the HeNB and HeNB GW architecture and Table 4.6.1-1 the X2 handover support, the definitions clause gives the hybrid cell, and the PCI split for CSG cells is quoted above.