Verizon’s first Draft Specification for LTE Devices

Unstrung has an interesting pointer to Verizon's first draft spec (v0.9) which details what they want devices to be capable of to allow them on their future LTE network. While they are mainly referring to relevant 3GPP specification documents, there are some nuggets of information in there that took my attention:

  • The 700 MHz band they intend to use seems to be band 13 in 3GPP talk. I wonder if that encompasses the complete 700 MHz band in the US or just a part for it!? In other words, will band 13 devices also be usable in other LTE networks (e.g. that of AT&T)? According to this post on Gigaohm, the total bandwidth of the 700 MHz that was auctioned off was 18 MHz per direction (uplink/downlink), so the answer is probably yes.
  • The channel bandwidth they will use: 10 MHz (double that of HSPA today but short of the theoretically possible 20 MHz as per the standard. Again, by looking at the link above, that's obvious because that's all there is available.
  • Devices must have RF connectors for testing purposes for all antennas. Haven't seen that in years on mobile devices. RF designers will have fun…
  • Devices will be assigned an IPv6 address when attaching to the network. IPv4 addresses shall be requested by the device if an application requests an IPv4 connection. The IPv4 address shall be released once no applications are executed on the device that require it. Interesting requirement, looks like an IPv4 address is not initially assigned to the mobile device by the network. This should not be a big deal, S60 for example already has a dual IPv4/IPv6 stack today.
  • Mobile device timer for moving from Connected to Idle state: The document says that the LTE standards say that the move from Connected to Idle state (on the radio layer, this has nothing to do with the IP address assignment) is controlled by the network. The standard leaves it open for the mobile to also initiate such a state change, for example if the device detects (by whatever means) that no applications currently wants to send and receive data. At this point, however, Verizon does not mandate devices to implement this. An interesting side note: Today, HSPA only knows a network initiated state change in the standards. In practice, however, there seem to be some devices that also trigger it from the devices side with a bit of an unorthodox signalling message exchange. Looks like standards people have learnt from that and included this feature in the LTE specs from day one.
  • No mention is made of dual mode CDMA/LTE capabilities. I wonder if that means that they expect that first devices will be LTE only? That wouldn't make a lot of sense to me. I can't imagine people would be very happy using a CDMA USB dongle and a separate LTE dongle, depending on where they are.
  • Verizon expects that first devices are data only, no voice capabilities. A pity, but who knows, they might yet discover the benefits of VOLGA.

Note that the current version is only a draft, there are still many unfinished chapters.

8 thoughts on “Verizon’s first Draft Specification for LTE Devices”

  1. Hi Martin, I suspect the reason why there is no multimode mention yet is that the Qualcomm chipsets will only start sampling in the middle of this year. I believe the only solution that is already sampling is Ericsson’s M700, which is LTE only; it can be paired with a HSPA baseband for multimode but presumably not a Qualcomm CDMA one.

    I suspect that is what they will start with and Qualcomm chipsets will come very shortly afterwards, maybe even before commercial deployment (or maybe not). Qualcomm will probably remain the only multimode baseband provider with CDMA for some time, although I’ve heard there might be at least one alternative eventually…

  2. BTW, I’ve got a question for you. I found your HSPA+ posts very interesting as well as the graphs you linked, so I was wondering if you might know how might know how frequent backhaul bottlenecked networks are compared to spectrum bottlenecked networks, and how this might affect the speed and/or usefulness of HSPA+ deployments.

    Clearly adding HSPA+ to handhelds in the 2H10 timeframe is feasible, but seems only justifable to me if it has a real bandwidth benefit and/or gets you more operator subsidies because of the higher spectrum efficiency. Do you believe either or both of those are likely?

  4. Hi Hyao,

    Thanks for your comment! I had a closer look and found the following:

    Verizon got the C-Block which is 746-757 (downlink) and 776-787 (uplink, give or take a MHz). That’s 3GPP band 13.

    AT&T won the B-Block and together with the spectrum they bought from Aloha wireless that should be somewhere around 704-716 (uplink) and 734-746 (downlink). As you say, that looks like 3GPP band 17.

    It’s interesting that down/uplink frequencies are flipped between these two bands!

    So I wonder whether it will be possible to manufacturer a single transcievers for band 13 + 17, or if Verizon and AT&T get their own hardware variants, or if it will economically be more feasible to manufacture devices with two transcievers.

    And of course on top, which other bands will 700 MHz devices support for LTE and other technologies to support national roaming to CDMA (Verizon) and HSPA(AT&T) networks and of course international roaming. I’d say that’s a formidable challenge.

    Cheers,
    Martin

    Two links for further background info:

    http://wireless.fcc.gov/auctions/data/bandplans/700MHzBandPlan.pdf

    http://en.wikipedia.org/wiki/United_States_2008_wireless_spectrum_auction

  5. Hi Arun,

    I guess it wouldn’t make much sense to upgrade to a higher speed on the air interface while leaving the backhaul as it is or vice versa. But I have no info if this is the case today anywhere. It’s a bit difficult to test as one would have to have at least 2 notebooks in each of the 3 sectors and, in case of Vodafone in the UK in cities, two per carrier per sector, which makes 12, to really test things. A bit beyond my possibilities 🙂

    Kind regards,
    Martin

  6. From the looks of the spec and comments on the VZ openDevelopment web board, Verizon won’t define requirements for multiple mode operation. And I don’t know how they could do that and live within the FCC requirements for open access and open devices.

  7. Verizon has left much to the imagination as might be expected.

    1 – They have to operate within the FCC open access and open device requirements for the AWS 700 C-Band. In response to questions raised following Verizon’s comments about how open they intend to make the network, the FCC said at the OpenMobile Summit last winter that they intend to enforce these requirements vigorously.

    2 – This likely means that VZ won’t require but cannot restrict VoIP. Seems to me that they could set guidelines to assure good operation but since they cannot control the revenue from VoIP on open devices, they might have little motivation to help out the competition to their own LTE-VoIP.

    3 – My understanding is that handoff between LTE and LTE, CDMA, 2G, 3G and even, heaven forsake Verizon, WiMAX, networks will be defined by the referenced 3GPP specifications. Since this is very important to user experience, particularly if used in a mixed service offering, VZ will probably test rigorously. But there is a question of how much VZ needs to offer to the open developer community. It is very certain that they will work with their network partners, Alcatel-Lucent and Ericsson and major handset providers for Verizon labeled devices and multiple mode service within the Verizon network. But they won’t likely have much motivation to spend money to enable developers to work out the kinks for other operators to hand-off and roam their LTE network.

    I’m doing creative speculation since some details remain unknown. I’ve asked VZ for details on many aspects of the development but they have not been overly forthcoming as of yet.

Comments are closed.