Short Term Weakness in Macro

The macro market is showing signs of vulnerability....despite the strength of end-user data demand. Component vendors, as always, are the canary in the coal mine here. A few component vendors (ranging from switches to isolators and filters) are reporting a slowdown in shipments over the past 2 months, which we expect to see in the announced results by public companies in the next few weeks.

Stay tuned. Overall we expect a minor adjustment with less than 5% decline in some areas during 1H2012, with strong growth continuing in LTE and a few areas of HSPA.

The Chips are Ready

Mobile operators would like to deploy small cells, but their reaction to some of the early products has been lukewarm. Picocells that are simply low-power versions of a high-performance macrocell are far too expensive. Femtocells designed for autonomous consumer use are suited to a coverage application, and do not play well in a dense urban network. Small cells mounted on a streetlight need expensive backhaul, which is generally not available.

These negatives are beginning to evaporate.

Recently, several SoC suppliers have introduced multi-core processors which are highly suited to work as a “base station on a chip”. The level of integration is awesome, with dozens of processor cores operating at high speed, with programmable DSP elements and hardware accelerators to handle high-speed computation while remaining very flexible.

The multi-core architecture is the key to success now. Simple reliance on Moore’s law to increase speed continuously is not enough: to make a small cell work, dozens of operations need to take place simultaneously, and each processor core must be suited to the tasks assigned. Several IC suppliers have recognized this need, resulting in strong products from Texas Instruments, Freescale, Mindspeed, Cavium, Picochip, Broadcom, and others. Each company has its strengths, but all of these “Base Station on a Chip” products utilizes a large number of high-performance processor cores and an internal interconnection fabric. Many of these alternatives can run the software from traditional macrocell or microcell products, making the integration with the macro network much easier.

The existence of 5-6 strong IC alternatives on the market will drive cost down quickly, and mobile operators will quickly start to see product offerings at acceptable cost levels….a few thousand dollars for a high-performance picocell or carrier femtocell.

The next question will quickly arise: What about backhaul? Depending on the capacity and bandwidth involved, some of the new ICs on the market can actually handle baseband processing for both radio access and backhaul….taking a major expense out of the equation. Mobile Experts is predicting a future where the “Base Station on a Chip” handles both datastreams, with RF transceivers branching in multiple directions. The result: Inexpensive, highly integrated small cells with multiple antennas, which can be deployed easily with licensed LTE operation to end users, and licensed TDD OFDM backhaul. The chips are coming together. Mobile operators need to start getting their backhaul spectrum licenses in order.

Mobile Transceivers

Transceivers for macro base stations are changing, as MIMO comes into the market. AAS will also have an impact in the 2013 timeframe.....but the overall trend between these two drivers is the reduction in power of a typical macro transceiver.

A 30W transmitter was the standard power level for GSM systems for 20+ years, and with CDMA and WCDMA systems there was a lengthy period of time where 40-50W was normal. Nowadays, however, the focus has shifted from coverage to capacity.....which means that operators are NOT looking to blast as much power as possible from a tall tower. Instead they are looking to transmit a measured amount of power, to place small cells and macro cells as closely as possible and re-use spectrum effectively.

This shift is apparent in the new Mobile Experts forecast which has come out in Excel form....see http://mobile-experts.net/product_info.php?products_id=33

Two Separate Small Cell Ecosystems

Looks like the small cell universe is splitting into two distinct ecosystems.

On one hand, we have Picochip and Broadcom, with their respective software stacks and APIs for femtocell vendors to develop consumer small cells. Taiwanese ODMs are jumping into this market because it looks a lot like the PC market or the smartphone game, where there are one or two dominant chipset vendors and an ODM can quickly wrap some plastic around it. In this market, the solution is a best-effort femtocell which can improve coverage in your home. If it doesn't interoperate with the network well (e.g. handoffs), that's ok! The fact is that you bought yourself a femtocell because you didn't have coverage at home.

On the other hand, we have Mindspeed, TI, Freescale, Cavium, DesignArt Networks, and the Tier One mobile infrastructure OEMs, all of whom are more focused on carrier-grade solutions. The silicon and the rest of the hardware will support the same software that runs on macrocell platforms, which allows the carriers to deploy a bunch of small cells with the level of service quality they already understand in a macro network. This ecosystem is NOT focused on coverage in the home. Instead they are focused on the more lucrative target: high capacity mobile networks used by thousands or millions of people instead of a few people at home.

Backhaul for Small Cells

The multi-core processors for small cells are coming to market....Freescale has introduced a great product, and Mindspeed has an excellent solution in customer hands already. But now the big question is: What about the backhaul? We can make a metro-femtocell for a few hundred dollars but do we have to spend $5,000 for a microwave backhaul link?

Dr. Jonathan Wells has joined Mobile Experts to investigate non-line of sight (NLOS) backhaul and its applicability to both femtocells and picocells. Our initial view is that NLOS will be perfect for some applications, but may not have the performance for advanced HetNet coordination.

What latency do you need in your application? What throughput do you need?

Weigh in on this topic with us....we are interested in inputs from our readers. Contact joe@mobile-experts.net or jonathan.wells@mobile-experts.net.

AT&T slows down on femtocells

AT&T has pulled back on deployment of femtocells, citing interference problems in areas with relatively high signal strength.

This is no surprise! Mobile Experts published a study in 2009 that indicated severe interference problems, even with femtocells that were "optimized" to avoid interference. So AT&T should not be so surprised that there are RF interference problems...

Our sources indicate that over the past six months, the AT&T deployment has dropped by more than 50% compared to the previous six months. They are now reserving their "Microcell" (femtocell) deployments to customers that have one-bar coverage in their homes. This is EXACTLY what we predicted and it shows that the consumer femto is truly a coverage solution only, and does not help in a congested network.

Femtocell shipments

Mobile Experts has just coordinated a worldwide survey to update our estimates for small-cell shipments. Looks like our February 2011 predictions were pretty close for this year so far...consumer femtos are shipping well, and we are seeing the beginning of the enterprise and operator markets.

Picocells are taking a little longer than expected to come to market...we will be looking into what's going on with the Tier 1 OEMs on this subject at the end of August.