Interview: Ericsson CEO on the rise of the HetNet


Credit: Ericsson

Lately we’ve written a lot about the heterogeneous network, or HetNet, exploring how today’s big-tower mobile grids will evolve into dense, multi-layered and tremendously high capacity networks. Given the complexity of such systems, it’s easy to imagine HetNet as a technology of the distant future, but the CEO of world’s largest mobile infrastructure maker Ericsson(s eric) doesn’t think so. In fact, in an interview with GigaOM on Wednesday, Hans Vestberg said the fundamental building blocks of HetNet are already here.

HetNets have three major components. The first is an umbrella — or macro — network designed to provide ubiquitous mobile broadband coverage. The second is a dense network of small cells that supply enormous quantities of bandwidth in the high-traffic areas its most needed. The final component is a network intelligence that ties those networks together. According to Vestberg, the wireless industry has largely built the first and is actively deploying the second. The network intelligence is still in the works, but it won’t be long before it’s a commercial reality, Vestberg added.

There are already 1.1 billion mobile broadband subscribers globally, traversing networks comprised of millions of macro cells, Vestberg said.  Only a small subset of them are LTE subscribers today, but LTE networks already cover more than 325 million people in North America and Asia.

“We’re at a tipping point,” Vestberg said. “You build the macro networks first, but we’re soon going to have very dense networks of small cells.”

The next step is Wi-Fi, which is already being deployed as a small cell technology globally. Major operators like Japan’s KDDI, Europe’s Orange(s fte) and AT&T(s t) are starting deploy dedicated high-capacity access points to offload cellular network traffic. Hotspots aren’t quite HetNet, but it’s the first step toward building a layer of small dense cells that can supply enormous capacity at low cost, Vestberg said. Ericsson has already gotten a jump on the competition by buying up metro-Wi-Fi vendor BelAir Networks, which supplies hotspot gear for AT&T and many U.S. cable operators.

Wi-Fi will soon be augmented by metrocells and microcells moving that high-bandwidth under-layer off of the unlicensed airwaves onto carriers’ own spectrum, Vestberg said. With those small cells they will be able to pack the total capacity of a macro cell into a tiny radius, and as those small cells multiply they will make once scarce mobile bandwidth plentiful.

The final component — the network intelligence to control it all — will deliver the true promise of HetNet, allowing operators to manage Wi-Fi access points as if they were any other cell and tightly integrate metro-and microcells into their network hierarchies without fear they will interfere with each other or the over-arching macro grid. That technology is already out of Ericsson’s labs and will be available in its forthcoming small cell portfolio, Vestberg said.

“I think North America will be one of the first ones to get there,” Vestberg said. “They’ve already started with Wi-Fi. They will do HetNet on top of that.”

Vestberg cautioned that HetNet is a gradual evolution of cellular topology, not a distinct network unto itself – a carrier won’t suddenly one day turn on a network of 1 million cells. The name “heterogeneous” provides a good clue on these networks will come together – Wi-Fi, home and public femtocells, pico-and micro-cells will be deployed in different stages but will eventually come together to form a larger network whole.



Re Michael Elling’s question, we fully accounted for the backhaul (and plenty of other) costs in the study at and small cells turn out to still play an important role, alongside spectrum and macrocell upgrades.

Dean Bubley


I think you will see some aggregation (& indeed already do) – although almost certainly not the sort of network-centric seamless approach that some envisage. Seams – borders, really – are too important to get rid of arbitrarily. See my blog post from a few weeks ago:

There are some use-cases where HetNets will be valuable, but equally there are plenty of others where that level of integration will be damaging to the user experience, or the desires & requirements of private WiFi owners.


Kevin Fitchard

Yeah, I see your point, Dean. You can’t really splice together a bunch of coffeeshop, hotel and airport networks, add a connection manager and call it HetNet. What about what KDDI is doing? The scope seems impressive.

Rupert Baines

One interesting aspect (and this touches on Dean’s “sep net” concept) is the degree the networks are integrated, and if so who is in control.

For example, Ericsson would advocate that macro, small cell & WiFi all come from one vendor: het net in layers, but “homo net” in service, vendor, OSS and purchasing (ahem)

An operatopr may prefer single network from services but the freedom to purchase from different suppliers: one can imagine Ruckus or ALU advocating this, as distinct from macro contracts.

Will .11u / HotSpot2.0 and the 3GPP SmallCell analogs support this – or is it a coherent, integrated, single vendor monolithy?

Kevin Fitchard

Hey Rupert,

Good points. I think ALU is firmly in the single vendor camp, though, as is NSN. The may not make Wi-Fi access points but they can easily partner with companies that do (or buy Ruckus). Otherwise they’re advocating their carriers adopt a single-vendor lifestyle. :)

The femto guys make an interesting case for a multi-vendor world, but their main argument seems to be low cost, which has merits, but if there is no integration it just won’t work. I think the femto guys would have us believe their cells can live in colonies without SON right under the macro network and somehow magically not interfering with the macro network.

As for Wi-Fi, great idea, but I’m shocked at just how hard it is to implement. Maybe it’s because there are three separate standards, three separate industry bodies trying to implement three separate technologies…

Rupert Baines


”femto guys” and small cell are esseentially the same.

The issue about just how well those devices live in a separate layer under the macro layer with independence & without problems is, indeed, at the heart of the debate.

For residential it clearly works.

The indications where small cell contracts are separate from macro contracts (for Korea with LTE, USA & UK with 3G) suggest that this is viable.

Dean Bubley

I think we’re more likely to get SepNets (separate networks) than HetNets. In particular, no more than a tiny fraction of WiFi access points will be controlled by (or partnered by) cellular carriers – in most countries, anyway.

There’s a lot of rhetoric about cellular networks treating WiFi just like “small cells” which totally overlooks the fact that WiFi has several more important use-cases than merely extending a service provider’s network. In particular, millions of locations provide free WiFi access for their own benefit (for guests & customers etc) and have no incentive to work with carriers on complex commercial partnerships.

Dean Bubley
Disruptive Analysis

Kevin Fitchard

I dunno Dean,

I agree that at 1,000 carrier-hotspot effort means little, but there are a lot of hotspot operators, telcos and and cable operators looking to sell their capacity off. If you can piece enough of those deals together, they could have a big impact. I’m not yet willing to concede that Wi-Fi offload is quite as silly as IMS. :)

Dave Wright

I think the mobile industry needs to back up and question some presumptions. Why does a successful Wi-Fi strategy need unified network intelligence between licensed and unlicensed? AT&T, O2, China Mobile, PCCW, etc… have all developed Wi-Fi architectures that are carrying significant portions of their wireless data traffic (54% for China Mobile and ~30% for AT&T). They did this without any unified anything, beyond an autoconnect feature on their handsets. Informa/Mobidia is indicating that 70% of global smartphone data usage may already be going over Wi-Fi (this includes non-operator Wi-Fi – home, business, other public, etc…)

While I’m goring sacred cows, they should also consider whether their efforts to make Public Wi-Fi access work similarly to, and in conjunction with, legacy BSS and OSS models is the right longterm decision. Those architectures were developed around a voice service paradigm decades ago. We could come up with much better/simpler solutions focused around mobile data today. If they don’t do it, some upstarts from the Wi-Fi side eventually will.

Operators who insist on this holy grail of a unified cellular/WLAN infrastructure are missing out on a huge market that is developing all around them as they dither. I don’t need my access networks to be unified and coordinated, I just need them to be available, to work, and to provide access to my services and data.

(Note, I think 802.11u and Hotspot 2.0 are great enhancements for Public Wi-Fi. But we need to see some architectures put forward that don’t involve integrating them with a legacy cellular core)

Kevin Fitchard

I see your point, Dave. It does seem an awful lot is done in the name of building a “carrier-class” network, when at the end of the day the consumer doesn’t really care. They just want Internet access.

Michael Elling

Call them hybrid access networks. The most important element will be the backhaul and cost of bandwidth. About 20x too high; if extent at all!

Comments are closed.