During Strong Angel II in 2004, Steve Birch, Steve Price and I conducted a series of experiments around a concept we called The Pony Express, which involved the use of Groove-based “mobile infrastructure” to deliver data synchronization services to users beyond the horizon of connectivity. For Strong Angel III, we will revisit the concept using the neutral, platform independent SSE replication protocol. First, I’d like to set the stage with a bit of background on the problem we were originally setting out to address.

The following scenario will be quite familiar to field practitioners experienced in humanitarian operations. A number of organizations tasked with supporting a response have deployed teams to a large disaster zone, far from headquarters. Members of diverse organizations are working in small, collaborative teams in isolated locations. These teams need to share information both with one another and with their respective headquarters offices. Headquarters, likewise, has information much needed by teams in the field. Local area networks, whether wired or wireless, are available, but they are underutilized. Wide area networks, such as satellite links with headquarters, are seldom adequate and are often unavailable. More often than most organizations would care to admit, their teams must operate in the field with no onsite network reach-back for days, weeks, and even months. Such users only have the opportunity to check email, browse the web, and use other network-based applications on the rare occasion when they drive to a location where email access is available. As a result of these communications challenges, most information sharing takes the form of face-to-face meetings. Such meetings require physical travel, often over great distances, incurring high costs, consuming time and fuel, and potentially making it necessary for staff to pass through high-risk areas.

The SA-II Pony Express was intended as a proof of concept that in environments where the communications infrastructure was compromised, the transportation infrastructure could be substituted. Rather than requiring that isolated users journey to the infrastructure, however, we wanted to explore the notion that the infrastructure itself could “go mobile” and deliver services out to the disconnected edge. For this series of experiments, an SUV was outfitted with capabilities that included a laptop-based Groove Relay Server and an 802.11 WiFi router. The concept was fairly simple. Members of the same Groove shared space (replete with files, discussion, calendar, and assessment forms) would be divided into teams, stationed in isolated locations many kilometers apart, and instructed to make a range of changes and contributions within the shared space. No Internet access was provided to users at any remote location. The Pony Express, under the command of Steve Birch, drove repeatedly in a circuit that passed within WiFi range of each group of isolated users. In each case, the users’ laptops attached to the mobile cloud and synchronized with the Groove Relay Server on board, uploading to it any changes they had made and downloading from it any changes stored on the Relay. The Pony Express would then drive on to the next site and the next, until it returned to “base camp” in Kona, where a satellite Internet connection was available. Once on the Internet, the Pony Express would synchronize all changes made by disconnected users with other shared space members located around the world. Having passed along the latest updates from the field, and having picked up the most recent postings from users elsewhere, the Pony Express would head back out to perform another circuit. An excellent composite slide of the Pony Express vehicle and components may be found here.

Through these experiments we were able to demonstrate the capability to keep isolated field teams relatively in synch not only with one another, but with remote users at headquarters – even when their laptops never had access to the Internet. Such concepts have been explored before; email servers and WiFi routers have been co-located in buses in India to deliver email to isolated villages. In the case of the SA-II Pony Express, we were attempting to demonstrate an analogous capability to synchronize structured data, such as evacuation requests and humanitarian needs assessments. While the experiments were moderately successful and did accomplish their goal, they were not particularly viable even as a starting point for implementation of a long-term practical solution – primarily because Groove clients were required for all users, and also because Groove users from different organizations would not, in a real world setting, ever be able to attach to the same Relay server.

For SA-III, we are taking an approach that we believe will be vastly more flexible and inclusive, and one which would be more readily developed into a practical, real-world solution. By way of context, recall that the scenario for SA-III assumes that the city of San Diego is under quarantine, limiting mobility, and Internet access is periodically unavailable. The Pony Express, in this case, will involve a laptop running IIS, SQL Server, and an SSE service. Two of these units will be stood up, and each will be co-located in a vehicle with adequate power and a mesh WiFi router. A range of SSE-enabled applications, running on various platforms and devices, will be configured to cross-subscribe directly with each of these SSE Mobile Relays, and each of the Relays will be cross-subscribed with each other. Users of SSE-enabled applications within each quarantine zone will create data structures such as Requests and Assessments. When a Mobile Relay comes within WiFi range, their devices will replicate with the Relay. The pair of Relays, having replicated with users distributed throughout their respective quarantine zones, will approach one another on either side of the quarantine boundary that they are not permitted to traverse. When in WiFi range of one another, the two WiFi clouds emitted from the vehicles will mesh across quarantine zones, and their cross-subscribed SSE Mobile Relays will replicate with one another. The two Relays may then return to their respective rounds, delivering to their users the Requests and Assessments collected in the opposite quarantine zone.The primary goal of this series of experiments is to demonstrate the ability to maintain shared situational awareness across quarantine zones, intermittent networks, organizational boundaries, and a mixture of devices, platforms, and applications. The team will be experimenting with a number of possible topologies and configurations, such as having all replication triggered by Relays, vs. by leaf nodes, and also with various polling intervals between leaf nodes and Relays and also between a pair of Relays.