Mesh app and service architecture (MASA)
Demand for apps that seamlessly stay connected as we move through our home, commute, and work are increasingly in demand.
“The purpose of a mesh network or app is that will it be high availability—everything connected to everything,” says Joseph Carson of Thycotic. “If the path is unavailable, it will find another device to establish the connection. We have seen this being used for example with the Tile tracker devices, which has created a community of tracking devices, and with bitcoin being a distributed ledger.”
But some see a lack of device compatibility as a potential bottleneck.
“Each vendor has their own way of trying to drive trust into this system, so they are all walled gardens, if they even exist at all,” says Derek Collison, formerly of Cloud Foundry and CEO of Apcera.
This technology promises a previously unthinkable level of connectedness—if a lack of standards doesn’t get in the way.
“My larger thought here is that AI will generally be trained in the cloud with massive amounts of data from all users,” says Collison. “These algorithms will then continuously update their execution model, which will be shipped to the edge over the air and update firmware on edge devices like our phones, cars, and home. The processing will happen at the edges in hardware; the training will happen in the cloud in software.”
Digital twins: Prepare to fail
Software models tied to physical and virtual sensors can help predict product or service failures so that organizations are able to plan and assign resources to make repairs before the failure occurs. Advances in machine learning and the adoption of IoT technology are helping to bring down costs for this sort of predictive “digital twin” modeling, which boosts efficiency and can bring down operating costs over the life of, say, a jet engine or a power plant.
Matias Woloski, CTO and co-founder of Auth0, says companies can also use digital twins in the concept and design stage, testing new products in simulations, then making changes until the engineers have the product they want. Findings from the digital twin are then used to build the product.
“A few organizations have already launched digital-twin initiatives, although the primary projects leveraging this technology are the ones with large upfront development expense where the cost of failure is too high,” Woloski says.
SpaceTime Insight’s CTO Paul Hofmann says digital twins benefit from machine learning, making them more effective than condition-based models at predicting failures.
“IoT and machine learning systems allow organizations to ensure that its assets aren’t randomly failing, and if they do fail, then organizations can optimize real-time decision making for the best long-term solution.“
Sign up for CIO Asia eNewsletters.