Posted by: nialljmcshane | June 12, 2010

Technology Wars: Will Public Cellular Win in Smart Grid?

In a webinar organized by Energy Central on Wednesday, June 9, Stephen Johnston, CEO of Smartsynch reprised the comments that he made at Greentech Media’s  Networked Grid conference that was held in Palm Springs CA on May 18-19 concerning his belief that public cellular networks would dominate the smart grid space.  This topic clearly attracts a lot of interest and it was reported that over 800 people were on the webinar representing public cellular carriers, utilities, regulators and Smart Grid network suppliers.

Johnston noted that utilities have been using public cellular for connecting to commercial and industrial customers for demand response for a long time and cellular is also used for backhaul of AMI data from aggregation points in mesh networks but addressed four main historical objections to the use of public cellular networks:

  • Cost: When Smartsynch started working with utilities in 2000, the cost per meter per month was around $15.  While this was acceptable for large commercial and industrial customers, who typically represent only 10% of customers but as much as 60-70% of the demand, it was prohibitive for the residential consumer market.  This price declined steadily over the years but was still sitting around $8 in 2009 when a dramatic change took place and the cellular operators realized that there was a large market to be accessed for little or no incremental network cost.  Prices dropped dramatically to around 50c per meter per month and are continuing to trend downwards.
  • Coverage: Smartsynch claim that 97% of the service area in the US is covered by cellular and, in a recent trial they were able to demonstrate 99.96% meter reading success rates.  This number actually reflects the percentage of the population within the service areas that are covered and also reflects an aggregate of all cellular providers rather than any specific network.  Another important factor when considering coverage metrics is that, unlike handheld cell phones that typically have a maximum transmit power of 0.6 W, fixed wireless devices such as smart meters can transmit at up to 2W which significantly extends the coverage area of any network that they are connected to.
  • Bandwidth: Concerns about missed data due to network congestion are a common reason cited for why utilities should build private networks.  However, Smartsynch  quote a statistic that, if all the water, gas and electric meters in the US, about 300 Million in total, were to transmit a day’s worth of 15 minute interval data every day, this would represent only 1/500th of 1% increase in the data volumes carried daily by a network such as AT&T.  They also note that, using quality of service settings, the cellular operators can prioritize utility traffic to guarantee delivery even when the network does become congested.  Johnston also pointed out that the cellular operators will continue to improve network bandwidth and reliability at no incremental cost to the utilities.
  • Security: Finally, on the topic of security, Smartsynch point out that the data transmitted over public cellular networks is not travelling over the public internet.  A significant amount of personal and financial data are transmitted over public networks which are NIST and DoD compliant and cellular operators continue to spend large sums of money on network security improvements.  Johnston claimed that the cellular architecture eliminates the multi-point vulnerability of many mesh network technologies and also pointed out that the newer IP security standards require higher bandwidth which the cellular networks can provide.

On the issue of total cost of ownership, Johnston acknowledged that the residential endpoints for connecting to the cellular network are approximately 10-30% more expensive than for competing technologies but noted that these endpoints have higher capabilities and intelligence than the endpoints used by other technologies and, unlike many competing technologies, there is no requirement for aggregation points or repeaters to complete the network connections.

In summary, Johnston claimed that

  • Cellular operators have a strong incentive to leverage their existing networks for Smart Grid communications.
  • Cost, coverage, bandwidth and security attributes of these networks are better than the alternatives and are getting better every day
  • Service level agreement with multi-billion dollar network companies represent reduced risk for utilities whose core competence is not in managing communications networks
  • Embracing open high bandwidth cellular networks will spur innovation by many vendors and not just a single vendor tied to a proprietary private network technology.

During the Q&A session that followed, a number of additional concerns were brought up and addressed:

  • In response to a question about the risk of technology rollover and meters being stranded, Johnston suggested that this could be mitigated by having a strong relationship between the utility and the chosen cellular operator and also noted that new technologies generally support backward compatibility with existing devices.  This is certainly true for devices that have a 10 year depreciation cycle but may not be true for equipment that is depreciated on a longer cycle as is often the case in the utility industry.
  • In relation to the important question relating to the fact that utilities in the US are allowed, by their regulators, to earn a rate of return on capital investment such as that required to build a private network but not on operating costs such as those paid to a cellular operator, Johnston pointed out that the recent FCC broadband plan recommended FERC to look at this rule so that utilities are not forced to make sub-optimal network choices.  The implication is that there could be a rule change in this area.  However, I would also note that even with a private network, the utility still incurs operational costs to maintain that network and those operational costs would also be subject to the rule that prohibits any earned rate of return.    At some point, the operational costs of paying the cellular bill will be lower than the costs of owning and maintaining a private network.  The point at which this occurs will vary depending on the price per meter per month for the cellular connection and the size of the utility’s private network.
  • In response to a question about reliability of cellular networks especially during emergency situations, Johnston noted that, during hurricane Katrina in New Orleans, the public cellular networks were among the last to fail and the first to be recovered.  He also pointed out that private network infrastructure was susceptible to destruction in such an event and would have to be replaced at the utility’s expense whereas any destruction of cellular network equipment would be replaced at the operators expense.

I invited representatives of several of Smartsynch’s competitors to comment on the claims that were being made in favor of public cellular networks but they either declined or stated that they were working on their own white papers to respond to these issues.  In part two of this article, I will provide feedback that I received from some parties who were willing to comment and will also discuss what this all means in terms of the overall Smart Grid deployment.

In the meantime, what are your views on this issue?  Do you agree with the arguments in favor of public cellular networks?  Are there other issues that were not addressed in this discussion?



  1. Smartsynch provides very powerful arguements here and I personally agree with all of their statements. I believe that public networks will win out – especially if the rule gets changed to earn a rate of return on specific operational costs. This is a game changer!


  2. Niall, excellent discussion!
    And I agree that the bandwidth won’t be a big issue at least for the foreseeable future. Sorry for my long winded analysis below:

    For an Automated Meter Infrastructure design on the West coast a few years ago, the amount of 15-minute interval data that would be transmitted once per day, during off-peak hours was estimated to be around 500 bytes for the data only.. (50 bytes per read) and about 2000 bytes when all the remaining packet overhead was included – per electric meter.
    This is not very much and for 10M meters it only required around 3.4Mbps of “Egress” bandwidth into the data center. Basically a little over two T-1 circuits.
    Over the next decade there will be huge changes
    a) when Demand / Response pricing systems are impemented
    b) “real time” requirements start to require multiple transmissions per day (x100 i.e. every 15 minutes as opposed to once/day)
    c) we start sending more information from gas meters, water heaters and Home Area Connected devices besides the information from the electric meters themselves (x4?)
    d) and the number of households with such systems increases to 100M (x10)

    There will be a geometric increase in transmission volumes but not overwhelming. The 3.33Mbps bandwidth would need to be (x100 x4 x10) or over 4000 times larger, over 13 Gbps. But that is still only a bit bigger than an OC-192. (around 10Gps)

    We are witnessing the birth of a new industry and it’s only a matter of time – like when the Internet started to grow in the early 80’s – before we see another leap in communications requirements. But it won’t be like the leap in Internet bandwidth from dial up to broadband per house. The bigger problem will be public acceptance and new applications, which could again turn this whole exercise on its head.

  3. Where I live in the UK, which, whilst rural, is hardly in the middle of nowhere, the signal for cellular mobile telephony is atrocious. So I have had to install a 3G femtocell, which somewhat ironically uses a VPN over my existing broadband (DSL). It’s a fair point that meters will probably be able to have a higher transmit power but the point is that there is never one technology entirely suitable for every situation. The cellular topology is dated in some respects compared to the dynamic and self-healing topology of mesh and the relatively low bandwidth required is also in contrast to cellular provision of higher bandwidth for the demands of the cellphone user.

    • Robert – I think you are correct that every technology will have some coverage gaps where additional or alternative technology is required.

      Note that mesh technologies rely on having other devices in relatively close proximity to create the dynamic, self-healing topology that you refer to. Can you comment on whether there are successful deployments of mesh technology in your area today?

  4. I’m not aware of any high profile usage of mesh AMI networks in the UK, but there aren’t many high profile AMI networks in the UK period, apart from the forthcoming British Gas rollout which admittedly uses cellular infrastructure for the AMI network.

    However, this is definitely not the case in the US, where Silver Spring Networks, Trilliant, Itron, Elster, Sensus etc. all have deployed systems using mesh technology.

    Others are betting on WiMax, some on BPL. Some see a mixed approach, for example, Itron have partnered with Verizon, giving mesh at the last mile and using cellular for backhaul.

    So, again, I don’t think there will necessarily be one dominant technology. What is important is that they are all based on the Internet Protocol (preferably IPv6), thus facilitating ubiquity and pervasiveness and enabling security through well-understood and widely deployed mechanisms

  5. Thanks Robert, are there any deployments of mesh network technologies for purposes other than AMI that would indicate the ability of such technologies to provide better coverage than cellular in that environment?

    • I think there have been some studies of, for example, muni WiFi-based mesh networks vs. cellular for the purposes of telephony and cellular tends to prove to be better in that particular case (although one paper I found was funded by Nokia and DoCoMo). But there are many factors which would render such a comparison in the case of smart grid communication of limited use. Even muni WiFi, although it tends to service the same sort of traffic as cellular, is aimed more at data traffic as opposed to voice, although the increasing use of VoIP may offset this. Smart Grid communications are typically quite different. AMI mesh networks are not typically the same as muni WiFi mesh networks, using different frequencies at longer range and lower bandwidth. Also, mesh is a relatively new technology and therefore it is not always easy to provide direct comparisons, especially in the AMI/Smart Grid space. One thing to bear in mind is where mesh originated from – military communications. The self-healing and dynamic nature of mesh networks (with appropriately deployed nodes of course) is a compelling reason to use it for critical infrastructure and the recent case of the Facebook mixup on AT&T’s wireless network shows these networks are not immune to security flaws which, when sharing traffic like this, could have significant consequences; a topic being carefully considered by the NIST SGIP cyber security working group.

  6. […] to the topic of Smart Grid Technology Wars, a reader sent me a link to a story on the local utility in Stratford Ontario turning the […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: