Managing the Internet of Things

Cloud-Based Remote Management Has High Impact on System Reliability

Secure, Cloud-based management agents are coming of age for the spectrum of embedded applications, enabling remote, centralized access to system data and dramatically reducing maintenance and management costs.


  • Page 1 of 1
    Bookmark and Share

Article Media

Embedded systems run the world. Whether the arena is transportation, defense, infotainment, medicine, communications or industrial automation, optimal technology performance is essential to the mission. Systems must be stable and reliable to run critical applications, often with non-stop, low-power performance while facing environmental demands like extended temperatures and extreme shock and vibration. Outages and downtime are not an option, and this imperative is a key performance requirement of connected systems. Managing a vital level of technology stability is a significant design challenge and requires advancements in control and system management tools that can detect potential problems before they become realized.

The shift to greater connectivity has immense impact on how system developers are handling this challenge. Today’s movement toward the Internet of Things (IoT), where smart devices share data in real time, has also driven embedded devices to evolve from isolated systems to connected, intelligent platforms. This offers a new era of system management and maintenance benefits that were not available for offline devices. System operators can now capitalize on Cloud access for centralized, proactive management that reduces costs by anticipating maintenance and avoiding location-based repairs. For example, rather than react to a system error after the fact, system operators can now remotely connect to view current system status, influence performance, and even predict, prevent, or troubleshoot critical system failures. Most importantly, this kind of access and value is available well beyond industries with an inherent need to monitor distributed networks, and can be readily implemented in nearly any embedded arena.

Better Decisions Reduce Costs

Knowing the condition of a system is an essential step in maintaining its reliability. With rich data at hand, operators can dispatch a service truck prior to system failure or manage the system to avoid a needless and costly service call. Eliminating errant service truck rolls is a boon to operators managing distributed networks—helping the bottom line and growing market share by keeping maintenance costs down and service up. Today, intelligent middleware tools—essentially a layer of software enabling remote management and analytics via a simple graphical user interface (GUI)—are facilitating this type of knowledge in real time. Operators can quickly address issues such as temperature increases and fluctuations in power consumption or fan speed.

For example, when a fan malfunctions, the system’s processor can overheat and become damaged. The system may go down, and repairs may be slowed by the challenges of replacing specialized components. When no replacements are available, critical systems may fail and costly downtime can extend from a few hours to a few days.

Using remote management, operators can ensure that repair personnel are dispatched to a system in distress before such a failure occurs. Systems can be reconfigured remotely to ensure they stay functional in the meantime. Operators can also capitalize more fully on remote management with ongoing analysis of system data. Long-term performance trends are revealed; system failure can be predicted and prevented in advance of any alarm signs; and system lifetime can be increased by monitoring and controlling various hardware parameters.

Integrating Remote Management into Embedded Solutions

Connecting to remote devices can be done in different ways, but all require hardware, firmware and software components. Adlink uses a dedicated board management controller (BMC), initially designed for power sequencing tasks. The BMC has since evolved to include many new and useful features for board management and control. Measuring the supply current to get a snapshot of the system’s power consumption is only one example of these new capabilities. And compatibility with the latest Embedded Application Programming Interface specification (EAPI) reduces design efforts to port existing calls to the BMC.

Providing the interface from the hardware to the operating system is one of the remote management system’s most important functions. The BMC first collects all relevant information from the chipset and other sources. Utilizing the System Management Bus, the application layer fetches the data and presents it to the user, displayed either in the BIOS menu or a user-friendly dashboard suitable for supervision and troubleshooting.

System operators can view and consider graphs illustrating various vital stats of the system, such as the power consumption or temperature of both the CPU and the board, queried every second and optionally written to a system log file stored locally for use by the system administrator. Data is written as plain ASCII text in tab delimited columns, allowing easy import into any spreadsheet type of program or other data processing tool. Users also have access to general board information, secure user access and storage areas, and fan, GPIO and I²C bus controls. The BMC uses smart fan controller technology, and automatically relates measured CPU temperature to fan speed.

Through their embedded board controller, local remote management agents also provide a defined amount of storage for normal end user data. This memory area is optimized to store serial numbers, keys, configuration data and other sensitive or board-specific information, as it remains independent from the BIOS and is not cleared or restored during BIOS updates.

A separate secure area provides additional storage, important for critical data such as secure key codes. This area can be protected through a one-time programmable hardware fuse to provide maximum security, and offers features similar to trusted platform modules (TPM) or SIM cards. System operators can attach a unique key to their system, preventing data from being read or copied without administrator permission.

Forensic information available after system or module failures includes minimum and maximum temperature of the CPU and system, as well as the cause of the last system restart event—all of which can be used to analyze system or module failure.

Adding Value with Device-to-Cloud Strategies

Extending remote management technologies to include secure Cloud access makes good business sense for system administrators. Cloud-based remote management enables always-available system data that can be used to increase system uptime by predicting and reacting to system failures or abnormalities without the time and cost associated with on-site maintenance of one or many distributed devices. As the Internet of Things proliferates, competition will increase. Providers will need real system intelligence that enables flexibility and a deeper understanding of system behavior under a variety of different processing loads and environments. This knowledge provides a competitive edge, keeping costs down, increasing system uptime and enabling smarter deployments in a greater range of creative new applications.

Cloud connectivity takes today’s intelligent middleware a step further than previous generations of remote management technology. By employing a Cloud server architecture and a machine-to-machine (M2M) stack on top of the intelligent middleware, embedded devices can connect to the Cloud without additional design requirements. Pushing data to the Cloud enables operators to verify, monitor and manage system performance from a single, central location—improving reliability and reducing management costs.

For example, the M2M stack in Adlink’s Smart Embedded Management Agent (SEMA) Cloud pushes system data to the user’s Cloud server via any kind of TCP/IP connection, such as 3G, LAN or wireless LAN (Figure 1). System managers have easy access to data and analytics through any commercial Cloud portal, using any device such as desktop PC, tablet or smartphone. 

Figure 1
Figure 1: Adlink‘s Smart Embedded Management Agent (SEMA) is a set of integrated, embedded functions, enabling Cloud-based, remote system monitoring and management. System operators can control various hardware parameters to increase the lifetime of embedded systems, and increase reliability through predictive maintenance.

Intelligent System Management

Using Cloud connectivity, three primary management scenarios cover the principal needs of system operators. These scenarios can be classified as information function, analytics and event creation, and multiple device management. Each plays a role in maximizing system availability, influencing performance by interacting with system data, and simplifying and reducing long-term maintenance.

Information Function: When systems are available, operators can observe their performance. Cloud-based remote management furthers that process by enabling observation anytime, anywhere. In this scenario, the embedded management agent continuously uploads data through an encrypted Transport Layer Security (TLS, the successor protocol of Secure Sockets Layer or SSL) connection, shown in the user’s information dashboard (Figure 2). The dashboard also shows temperature and power consumption information for different parts of the embedded system. Since data can be accessed at any time, operators can determine if performance is acceptable even if certain values fluctuate from normal settings.

Figure 2
Figure 2: Using the remote agent’s user-friendly dashboard, limits can be defined for several types of system data. When thresholds are met, operators are alerted quickly via text or email.

Analytics and Event Creation: The same user-friendly dashboard allows system operators to define limits for several kinds of data. In this case, the Cloud application software continuously analyzes the incoming data and, if user-defined limits are reached, an alarm will be issued. Using a mobile device with two battery packs as an example, the device is running from the primary battery with a secondary battery as backup. The backup battery becomes active if the power capacity drops below 10 percent as reported by the embedded management agent monitoring power consumption.

If the capacity of the primary battery pack drops below 10 percent, an alarm is generated and the remote agent switches instantly to the secondary battery pack. In parallel, the system operator is informed via SMS text message or email that the device must be charged.

Operators can proactively interact—rather than simply react—with the system for better reliability, dealing with potential issues in advance as well as responding quickly to downed systems. Consider, for example, the ice freezer at a gas station: the ice is stored onsite but belongs to an offsite ice vendor and is valued at thousands of dollars. Temperature issues can quickly dissolve these assets. However, vendors are implementing M2M monitoring devices that alert managers immediately to issues and ideally prevent damage and financial loss before it happens.

 Multiple Device Management: Cloud-enabled embedded agents offer the additional advantage of being able to remotely control system parameters; specific user configurations will trigger certain actions to execute automatically (Figure 3). This is possible for a large number of devices, enabling a form of fleet or multiple device management. Through the M2M Stack, users can easily set up a Cloud application to control different devices, and the Cloud application observes the current health status of the connected embedded systems. Before a device fails, the Cloud application may recognize the malfunction through these remote management functions, enabling a quick reaction, such as shutting down a system before any harm comes to it. System operators have the advantage of being able to rescue systems, as well as check and correct malfunctions. Repair costs are reduced, and workloads can be redirected from one system to another, which eliminates downtime in event of equipment failure.

Figure 3
Figure 3: Operators can remotely control system parameters such as fan speed; actions are triggered automatically based on system health and performance, preventing system damage in case of malfunction.

Further, system longevity is increased when administrators can react before severe hardware damage occurs. A rising CPU temperature illustrates this concept and shows how remote device management intervenes. The BMC uploads the data to the agent, which reacts immediately by attempting to increase fan speed. If this is unsuccessful due to hardware failure, the system is remotely shut down for safety, and the operator is notified simultaneously. Once notified, the system manager can replace the fan and quickly restart the system.

Broad Promise as M2M Service

A centralized, Cloud-based approach lends itself well to an intelligent services business model, where system operators subscribe to the level of monitoring and management appropriate for their application and number of devices (Figure 4). As M2M strategies unfold—for example in healthcare, smart metering, smart homes, POS and retail banking, factory floor systems and connected buildings—the business case for intelligent services increases.  A recent Juniper Research report forecasts that M2M service revenues will reach $20 billion globally in 2015, fueled by manufacturers and developers simplifying the process of rolling out secure M2M strategies for the end user.

Figure 4
Figure 4: : Applications such as fleet management, public safety, utility substation monitoring or any implementation that encompasses field service or a broad distributed network, are poised to embrace advancements in remote access and management via Cloud-based architecture.

The transaction and access of sensitive data using Cloud-based M2M applications requires security considerations at each level of the solution: at the device, during data transmission and in the Cloud environment. On the device level, software-based control tools such as whitelisting can be used to protect locally derived and stored data. As previously mentioned, encrypted protocols like TSL cover the connection between distributed devices and their Cloud-based data access points. In the Cloud, hosting companies have an arsenal of software tools and encryption methodologies available to protect data residing on virtualized servers.

Certain “classic” embedded applications have an intrinsic investment in remote management services. Applications such as fleet management, public safety, utility substation monitoring, or any implementation that encompasses field service or a broad distributed network, are poised to embrace advancements in remote access and management. However, Cloud-based management services have a much broader application, opening doors in the spectrum of embedded arenas. Industries where remote management would have been an asset, rather than a requirement, now have easier access to sophisticated data that enables a tangible competitive edge.

Systems may connect broadly using 3G wireless; they may also simply reside in a factory’s central office, receiving data via wired or wireless Internet connection from systems throughout the factory floor. Cloud access is available in solutions that start at quantity one, enabling all manner of commercial goods manufacturing to benefit from tapping into powerful, real-time system data. Medical devices, industrial automation, office equipment, or devices that are fixed installed or moving in the field—any application that incorporates an embedded board—is a candidate for Cloud-based remote monitoring services. These tools and services may also include the capability for remote software and OS updates, allowing users to easily update firmware and upgrade the BIOS over-the-air, adding advanced features and pushing them out to devices in the field. 

Embedded, connected systems can generate and collect a vast amount of system performance data—and designers are now capitalizing on Cloud technology to share this data for reduced costs and improved reliability. Remote management eliminates the need for proximity to maintain and troubleshoot distributed devices; this reduces costs associated with both physical travel and system downtime. In addition, Cloud access means critical systems are available for observation from a simple, centralized location; operators can remain informed about system health and status, and use real-time system intelligence to make better, more cost-effective decisions regarding service and performance. Service is more strategic as administrators influence and interact with system performance, predicting and preventing failures in advance of critical alarms.

Information gathered, shared and applied creates a better end-user experience, reduces costs and builds revenue, enables new applications and improves the overall value of technology.  Automating and streamlining these benefits through remote monitoring and management is a supremely practical application of M2M technology, as any embedded market benefits from detecting potential problems before they become realized.

ADLINK Technology
San Jose, CA
(408) 360-0200