With Sigfox and LoRa the current big-names in the Low Power Wide Area Network (LPWAN) space, companies looking to embrace the IoT networking technologies can now (fairly simply) go out and set one up. As it stands, doing so would currently tie that customer into the ecosystem of one the two leaders, as both technologies use proprietary IP.
For many businesses, this isn’t an issue in the slightest – it’s almost a fact of life. But there will always be a market for the open standard implementation, and while it might not be as big as the proprietary alternatives (think of Linux vs. Windows in terms of PC share), the open standards often fill a practical, and sometimes political, purpose.
Last week, the Weightless SIG announced the pre-standard version of Weightless-P. The new open standard aims to address the longer ranges required in some networks, and act as an augmentation to its Weightless-N protocol – which is much better suited to traffic that is predominantly flowing from the network-edge to the cloud. With the two Weightless protocols, the SIG is offering a direct challenge to the likes of Sigfox and LoRa – make no mistake about it.
This week, we spoke to Haystack Technologies’ CEO Patrick Burns, who told us about another open LPWAN challenger – DASH7. Based on the ISO 18000-7 Active RFID standard, from where it derives its name, DASH7 operates in the license-exempt 433MHz band, with FSK modulation. Haystack is the author of the spec, and has received strong interest from the mobile, agriculture, consumer electronics and automotive industries.
DASH7’s most notable difference to the other LPWAN implementations is that it defines the full OSI stack, and can be adapted for use with other PHYs – such as Sigfox or LoRa, and Burns said that it has had interest from potential customers looking to do such a thing, meaning that we could begin to see deployments using Semtech silicon, an Actility mid-stack, and DASH7 in the upper elements.
Another very noteworthy aspect of DASH7 is that through its use of the 433.92MHz band, thanks to the proportional relationship between that frequency and NFC’s 13.56MHz (433.92/13.56 = 32 or 2^5), the same NFC radios can be used in DASH7 devices, meaning that an NFC device can use both NFC and DASH7 – by temporarily tweaking the NFC antenna and using harmonics to reach the 433.92MHz band.
This gives DASH7 a potential route to market via the millions of NFC-compatible smartphones and devices that are sold each year, as long as it can be implemented in software. Reflecting on the adoption of NFC, Burns thought that Apple was ready to launch its NFC platform some two years earlier than it did, and that now it looked more like Android would beat it to DASH7 adoption – although whether the stack is passed down the chain from Google to the carriers and on to the users is another matter.
That fragmentation could harm adoption, and it’s something Google is painfully aware of as it struggles to roll out the latest versions of Android and patch gaping security holes, but the fact that DASH7 can use readily available silicon from multiple vendors is a big plus.
It can also communicate on a tag-to-tag basis, meaning that it can be implemented in labels and communicate with other labels or tags without passing through a reader or base station. This gives it capabilities similar to mesh networks, like ZigBee and Z-Wave, where devices are able to communicate at much longer ranges than their antenna should allow by means of hopping between different end-points.
On that point, the range that most DASH7 documents claim is 1km, but ranges up to 10km have been confirmed – typically in the EU, which has more favorable spectrum regulations that the FCC, something that Burns mentioned in our conversation. That range is much larger than the likes of ZigBee, but also considerably less than the maximum ranges quoted by Sigfox and LoRa.
The protocol also allows for a gateway to dynamically query the devices in proximity, rather than the process used in other LPWAN protocols where end-devices report to the gateway at predetermined intervals. This function lends itself well to asset tracking, but is not such a priority in fixed-asset deployments. Haystack says its DASH7 implementation has a best-case latency of 1 second, and a worst-case data latency of 10 seconds.
The queries themselves can be structured to avoid receiving unwanted responses from nodes. For example, a warehouse gateway could ask to find all containers that are destined for Europe, and receiving only positive acknowledgements. If only 10% of the warehouse is EU-bound, then the process would prevent 90% of the other containers from generating network traffic – but you would need to include a process to ensure that every container received the query, or simply run it multiple times to ensure so.
The DASH7 stack also includes an integrated file system, engineered for the kinds of sub-$1 MCUs that the protocol will likely find itself on. This allows the MCU to store data or events on the chip, and send the results in response to a query, rather than have to send them out as they are logged. It also allows the sensors to implement queries such as “send me a notification whenever a 3+ year old cow moves on the farm,” for better clarity and reduced network traffic – in which all cow movements would be logged, and then the age requirement query implemented in the cloud installation.
Haystack and DASH7 have a history in military supply chain management, being used by Haystack’s predecessor Savi Technology (now owned by Lockheed Martin) as part of a 2009 deal to monitor the position of assets inside the US Department of Defense’s logistics processes. DASH7 is still being used by the DoD today, as well as NATO.
Using the RFID protocol in natural chokepoints, like warehouses or vehicles, the contract allowed Savi to locate shipping containers in the field. Shortly after that contract was signed, the DASH7 Alliance was formed, with the non-profit aiming to promote interoperability between the devices. The protocol was open sourced to non-members in March 2012, although Burns notes that despite being a founding member, Haystack itself doesn’t prioritize the alliance as it isn’t looking to spend lots of its time working with an industry association.
Haystack is more focused on smaller developers than militaries these days, and builds tools, APIs and software to help those developers get a foot in the DASH7 door.
Burns also noted that the LPWAN space is a hard one to navigate for adopters. Picking a company to facilitate a ten-year deployment, in which the business model might only make financial sense if it is supported for such a length of time, requires faith that the company you choose will still exist in a decade. Similarly, opting for one of the PHYs ties you to that choice, as the LPWAN PHYs are all rather different.