|By Kathy Lee Simunich, Gordon Lurie, Michelle Kehrer, Tom Taxon||
|March 5, 2004 12:00 AM EST||
The need for a chemical/biological early warning system within mass transportation sites is real and was being researched before the September 11 terrorist attacks. Since then, this research has become an operational necessity.
Led by Argonne National Laboratory, along with teams from Sandia National Laboratories (SNL) and Lawrence Livermore National Laboratory (LLNL), researchers created an early warning system for subways within a major metropolitan area. The entire system entailed the processing of sensor data, live meteorological data, video feeds, and real-time track/train data; execution of below- and aboveground dispersion models; and the timely display of results to subway and emergency personnel. The system is based extensively on Java technologies, and this article examines the merits of each J2EE package used.
The Chem-Bio Emergency Management Information System (CB-EMIS) is an enterprise-level distributed system. The decision to write it in Java was made early because of Java's maturity, scalability, and maintainability, as well as the availability of the industry-supported J2EE technologies. This article discusses the software aspects of the CB-EMIS and presents the architecture in detail.
Industry Standard Technologies
The CB-EMIS runs on four dual-processor Dell PCs running the Red Hat 8 Linux Operating System (the servers). Because the system must always be up and able to recover from communication failures, etc., three J2EE technologies were used to gain this reliability: Java Messaging Service (JMS), Remote Method Invocation Object Activation Daemon (RMID), and persistence through Java Database Connectivity (JDBC). A relational database was used to persist data for forensic playback and simulation modes (MySQL for development and training, and Oracle 9i for operational use).
The Sun ONE (previously iPlanet) Message Broker was used for the JMS system, and Novell's Lightweight Directory Access Protocol (LDAP) implementation (eDirectory) was set up for use as the directory service and for the authentication of users. It's important that the server processes across all CPUs remain synchronized in time; a Network Time Protocol (NTP) server was used to keep the system times in sync. Since the system relied on JMS messages, it was important to have the times match for message time-out purposes. Also, the Time Server process reads the system time to keep a consistent time stamp for all data and messages passing through the other server processes.
CB-EMIS System-Level Services
This section describes the design of the foundational services layer with which all server processes extend, implement, or interface. Figure 1 shows the services layer that supports the CB-EMIS.
The LDAP server was used for authentication of user accounts as well as a directory service for object binding. The schema was also extended to allow the storage of attributes unique to the system on a per-user basis, allowing control over what each user could see or do in the client graphical user interface (GUI). The Java Naming and Directory Interface (JNDI) package was used throughout the CB-EMIS to bind and look up object references stored in the LDAP server. Figure 2 shows the various objects bound into the directory. Not only were the JMS Topics and Topic Connection Factory references bound in the server, but the various JDBC data sources were as well. In this way, any of the CB-EMIS processes can look up the references to all the connection factories, topics, and JDBC data sources without having provider-specific references within the code. Saving these objects within an LDAP server gains persistence and security, and avoids the need for a shared file system. The team also decided to bind the CB-EMIS server process Remote Method Invocation (RMI) stub references within the LDAP server. Because there's not a JNDI class for binding stub references within a directory server, it was necessary to write our own javax.naming.spi.ObjectFactory and javax.naming.Referenceable classes to serialize/deserialize the java.rmi.server.RemoteStubs.
A relational database management system (RDBMS) is used as the persistent store (referred to as the Archive) for the static and dynamic data during either a simulation or normal operation. The purpose of persistence is for error recovery and playback mode. Static system description data is imported into the database for use during normal operation or simulations. During playback mode of either a training simulation or a normal operational session, the data archived in the database is retrieved and executed for the user. Archived live data can be played back for forensic study of an incident.
Each server process looks up the data source via JNDI and connects to the appropriate database through the JDBC. All events and data are saved in the database for playback or for error recovery. Rather than having SQL statements scattered throughout the code, a Persistent Object class was created that other classes extend. By defining the table name and the column names, each class inherits the ability to insert, retrieve, update, and delete itself from the database. JDBC Prepared Statements were used to speed up transactions to the database.
A Persistence Broker class does the JNDI look-up, maintains the reference to the data source, and automatically reconnects to the database if the connection is lost. Most of the data records are time-stamped, which is usually part of the primary key.
Activatable Server Processes
Each of the server processes is restartable via RMI activation. Activation allows programs to register information about remote objects so they can be started up when they are first accessed or when the RMI Daemon (rmid) starts up. This provides the ability to start the entire server, which consists of processes across multiple machines by registering one process to be started when rmid starts up, as well as the ability to recover from a Java Virtual Machine (JVM) failure by restarting a failed process. If the JVM of an activatable object does not exist, the RMI Daemon creates a new JVM for that object.
All of the CB-EMIS server processes are restartable by extending the java.rmi.activation.Activatable class. The process stubs are registered with the rmid process (which is already running) and also binds its stub reference into the LDAP server. The RMI Daemon provides a Java Virtual Machine from which other JVM instances may be spawned. This process actually starts up the other server processes "on-demand" as they are requested. Therefore, if a process goes down it will be reactivated upon the next reference to it via rmid (barring an actual hardware failure).
A stub reference is looked up by a CB-EMIS process via JNDI and passes that reference to the RMI Daemon, which in turn activates the process.
All the server processes communicate via a publish/subscribe methodology with JMS Topics. This communication mechanism was chosen because JMS has several advantages over Remote Method Invocation for this type of application, most notably:
- It is asynchronous in nature, so if one component goes down, the process that's sending the data is not tied up waiting for the other process to restart.
- Messages can be guaranteed to be delivered even if the process that needs it is not running. It will get all waiting messages when it starts back up.
- It's fairly simple to have another process register to receive data from the same source as others without having to contact the source of the data.
CB-EMIS Server Processes
Figure 1 shows the CB-EMIS Server Processes interacting with the services layer. The CB-EMIS "server" has been broken down into the various components (or processes) described in this article based on the various tasks the server needs to do.
The processes fall into the following categories: data collection, monitoring, and utilities. The data collection processes are the Meteorological Data Collector, Live Met Data Feed, Track Data Collector, Sensor Data Collector, and the Critical Actions Collector. The monitoring processes are the Alarm Generator, System State Updater, and System Monitor. The utility processes are the Time Server and the Model Executor.
Thus, the server consists of 10 separate processes running within their own JVM across the four dual-processor PCs. They all are activatable via the RMI Daemon. When they start up, they make a JNDI look-up to get references to the JDBC data source and make a connection to the database as well as to the JMS Broker to sign up for publishing and/or subscribing to the available topics. All of the processes employ multithreading to provide the fastest possible processing of information through the system, while monitoring all the data topics and archiving to the database.
This section describes each server process, its basic purpose, and the relevant Java technology it employs to do its work. As stated earlier, the processes communicate via JMS Topics. These include:
- System Control: The System Monitor publishes to this topic whenever the administrator changes the mode of the server (normal, playback, or training). All processes subscribe to it and change their processing according to the current mode of operation.
- Process Status: The System Monitor also publishes to this topic approximately every two seconds so that the administrator GUI can show the current state of the server processes.
- New Data: All server processes archive data and events to the database and then publish the data on the New Data topic for the other processes to ingest (if interested in the particular data item).
- Critical Actions: The CB-EMIS client process publishes to this topic whenever the emergency response users declare a critical action (e.g., trains have stopped, alarms triggered). The purpose of this topic is to provide a non-RMI method of having the client processes communicate critical information back to the server processes.
- System State: The System State Updater publishes eight different data types to this topic. Each data type runs in its own thread, publishing frequently enough to update GUI elements within the client process. By publishing the state of the system onto this topic, any number of client processes can log in and display the current state by subscribing to the topic. As a consequence, clients can also recover from communications interruptions without missing data.
The Time Server
The Time Server is responsible for holding and maintaining the system time. It's the central point for all the server processes to obtain the synchronized system time, since it is the process that is synchronized with the NTP server. The other server processes access the time from the Time Server process via an RMI call. Having a single process keep track of the system time allows all processes to use the same methodology for getting current time regardless of the current mode (normal, training, or playback).
The System Monitor
The System Monitor is responsible for monitoring the state of the software system. It restarts the components if there is a problem and notifies the user if it cannot restart a process. The status of the system processes (components) is displayed via a GUI for the system administrator.
The System Monitor is responsible for verifying that all required processes are running and maintaining the current run state (mode, simulation run status, simulation speed, etc.). It pings a process via RMI every two seconds. When a process is pinged, the activation system (RMI Daemon) starts up the process if it is not already running. The individual processes must also ping the system monitor back within five seconds. This allows the administrator to be notified if a process starts up and then crashes immediately. In addition, if a process detects or recovers from a nonfatal error, it can notify the System Monitor of this state change via an RMI call. If the status of a process changes, the current status is published on the Process Status Topic.
The current run state is maintained via RMI calls from the System Monitor GUI. When changes are made to the state, they are published to the System Control Topic so that all other interested processes can handle the change.
System State Updater
The System State Updater is responsible for maintaining the system state (train location, track status, meteorological conditions, sensors, plumes, alarms, etc.) and publishing it periodically to the System State Topic. There is currently one updater for high-bandwidth connections to the server, as there is a full, detailed system state sent every second. It listens to the New Data topic for when any process saves data to the database and then publishes the new data on the appropriate System State Topic for consumption by client processes that are currently connected. The design allows for other levels of detail of the System State to be published. One such use of this would be a handheld computer used by an emergency responder; it might get only system state data relative to the responder's location. Thus, a low-bandwidth updater can be created that publishes updates less frequently for clients connecting over a dialup or wireless connection.
The Alarm Generator is responsible for monitoring the New Data Topic for sensor events, determining if an alarm condition has been reached, and storing the data in the Archive/DB. It also publishes any alarms to the New Data Topic.
This process is responsible for responding to sensor alarms in the New Data Topic. It generates input files based on data received from the New Data Topic and starts to execute the appropriate below- and aboveground dispersion model. When the model(s) are finished, this process publishes the contamination results to the New Data Topic and archives the data in the database.
Critical Actions Collector
The main purpose of this process is to provide a non-RMI method of having the client process(es) communicate with the server. This is essentially a pass-through for Critical Action Data messages. It tracks who created a critical action and when it was generated.
Sensor Data Collector
This process is responsible for monitoring all the chemical/biological agent sensor readings. It stores the readings in the Archive and publishes the data on the New Data Topic. This process registers with the SNL Sensor Server via RMI.
Live Met Data Feed
This process polls the LLNL Met Server, which monitors real-time weather observations, and stores the data in a local database for access by the Meteorological Data Collector.
Meteorological Data Collector
The Met Data Collector is responsible for retrieving meteorological data from a server (or from a file if in training mode), storing the data in the Archive/DB, and publishing to the New Data Topic. This data is needed for the dispersion models to correctly predict plume concentrations.
Track Data Collector
The Track Data Collector is responsible for receiving real-time track status data and train location data, archiving it in the database, and publishing the data on the New Data Topic. It subscribes to the Critical Actions Topic for notification if the trains stop running.
CB-EMIS Interactions with Other Hardware/Software
Figure 3 shows the CB-EMIS interactions with various servers that have the physical connections to the sensors as well as other parties' monitoring software with which the CB-EMIS must communicate. Four types of sensor hardware are needed for the system:
- Chemical/biological agent detectors
- Meteorological sensors
- Real-time train location and track status sensors
- Real-time video camera feeds
Argonne has written a belowground dispersion model specifically for the subway where the CB-EMIS has been installed. The system also employed an existing EPA aboveground model (INPUFF) for the prediction of plumes if the agent escapes from underground. These models are written in Fortran and therefore use flat files for input and output. The CB-EMIS Model Executor reads/writes the output/input and makes a system call to actually start the models running on a separate CPU.
The CB-EMIS also uses the LLNL real-time meteorological observation server. Data is updated every 15 minutes, and the Live Met Data Feed process polls the server for current observations. The Model Executor can also access and run LLNL's high-resolution aboveground dispersion model (Narac) at a preset time after an alarm to obtain more accurate results. The default INPUFF model is useful for quick assessments, but more detailed predictions can be calculated with the Narac model (although it takes longer to run and get the results).
Sandia was responsible for installing the sensor detection software interface to log the actual readings of the chemical/ biological agent detectors. They implemented their sensor collection software as a Java servlet and, therefore, the servlet must run within a servlet container. The Apache Tomcat Server is used as the servlet container and Web server and runs on one of the nodes in the Linux cluster. The servlet continuously monitors the sensors.
Figure 4 shows one view of the Swing-based GUI for the emergency response coordinators. Each panel on the display can be either minimized and shown in icon form or maximized to show more detail. Panels include:
- GIS-like overhead map view with zoom and pan capability
- Overview map of subway system and train status
- Plume concentration overlay
- Sensor concentration-level graphs
- Critical actions and alarms
- Video camera views
Emergency response systems, such as the Chem/Bio Emergency Management Information System, should be deployed throughout many critical infrastructures such as subway systems, airports, or other mass-transportation hubs. This article presented the distributed processes, the need for asynchronous communication through messaging, and secure communication and user access mechanisms provided by the CB-EMIS employing various Java technologies.
Reliability was achieved by using the RMI Activation Daemon to reactivate any processes that may have failed, and persistence of JDBC Data Sources and JMS Topic Factories within an LDAP server allowed restarted processes to easily look up the references via JNDI and reconnect to the database and message broker automatically.
Security was maintained through user authentication within the LDAP server on both the CB-EMIS server side as well as access to the client processes. RMI security policy files were used to restrict access rights for any RMI calls throughout the server processes. Further security was achieved by using an SSL connection via JSSE from client processes to a proxy server and then to the server processes behind a firewall.
This project has proven that Java technologies not only addressed the various needs of the system, but helped all the components of the system integrate seamlessly and efficiently. With all the monitoring processes, the continuous real-time updates (not only of the system status, but of the train and track data and sensor readings), the database archiving of all data, and message publishing and subscribing, the performance of the multithreaded processes running in multiple Java Virtual Machines was never a problem.
Q & A
Joe Ottinger, JDJ's editor-in-chief, had the opportunity to have a discussion about the Chem-Bio Emergency Management Information System with the team at Argonne National Laboratory.
JDJ: What other technologies were considered? What were the measurements for validity for the technologies being considered?
ANL: We wanted to use the established Java technologies. The first principle of the entire system was reliability.
The early prototype system was a Smalltalk/Java program using RMI and some custom socket-based communication between Smalltalk and Java. The Smalltalk version of GeoViewer was existing code that is now converted to Java for the final product.
Once we converted to production development, we needed to manage the synchronization of the communications between all the server processes and clients, and each process had to be restartable. We chose RMID activation for its restarting ability and JMS for its asynchronous communications and guaranteed delivery of the messages.
These technologies greatly reduced our development time.
JDJ: I notice you're duplicating a lot of the functionality of JMX - how has the JCP worked to affect your development process and architecture (why wasn't it used, in other words)?
ANL: We felt JMX was not a mature enough technology when we started on the final system. We started the final development stage in December 2001 and it had to be delivered/ installed by January 2003 and operational by June 2003.
JDJ: What licenses did you use for your custom code, specifically for the RMI stubs in your LDAP server?
ANL: We used LDAP for authentication and the JNDI lookup of the data sources as well as the RMI stub references. The use of eDirectory was dictated by the sponsor.
JDJ: What kind of persistence layers did you consider, if any? Are they in-house, commercial, or open source?
ANL: We had already written an in-house persistence layer for other projects, so we adapted it for use in our final prototype. The database was used more as a logger than needing the full features of an RDBMS. We used generic SQL to switch between using MySQL for development and Oracle for deployment.
JDJ: You mention you're using iPlanet for some services - why, then, did you choose Tomcat for a servlet container, when iPlanet also has a servlet container available for it?
ANL: We bought just the Sun ONE Java Message Queue since that's all we needed it for. We chose it after testing a couple of other vendors' products. The servlet used to monitor the sensors was a prototype developed by another party that has since been replaced by a more robust system, which is not based on servlet technology. It's much more tightly integrated with the rest of the system.
JDJ: In the "System State Updater" section, what is a "plume"? I have some idea because of context, but clarification would be nice.
ANL: The "plume" is a contour plot of the predicted concentration levels of the released agent that could pose health risks.
The plumes are updated every minute and are plotted as an overlay over the map on the client machines.
JDJ: What client platforms are supported?
ANL: Windows 2000 and XP have been extensively tested for the client software. However, plans are in the works for supporting Red Hat Linux (which is what we run on the server side). All the code runs on the major OSes, including Solaris, Red Hat Linux, and Mac OS X. Basically, any platform that supports J2SE can support our system. There is a small amount of JNI code included in our system, so the biggest hurdle is just compiling that code for a new platform. Plus, tweaking for optimal bandwidth usage for the video streaming needs to be done on any new platform.
The submitted manuscript has been created by the University of Chicago as Operator of Argonne National Laboratory ("Argonne") under Contract No. W-31-109-ENG-38 with the U.S. Department of Energy. The U.S. Government retains for itself, and others acting on its behalf, a paid-up, nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, by or on behalf of the Government.
Manufacturers are embracing the Industrial Internet the same way consumers are leveraging Fitbits – to improve overall health and wellness. Both can provide consistent measurement, visibility, and suggest performance improvements customized to help reach goals. Fitbit users can view real-time data and make adjustments to increase their activity. In his session at @ThingsExpo, Mark Bernardo Professional Services Leader, Americas, at GE Digital, will discuss how leveraging the Industrial Interne...
Apr. 29, 2016 10:00 AM EDT Reads: 1,091
Whether your IoT service is connecting cars, homes, appliances, wearable, cameras or other devices, one question hangs in the balance – how do you actually make money from this service? The ability to turn your IoT service into profit requires the ability to create a monetization strategy that is flexible, scalable and working for you in real-time. It must be a transparent, smoothly implemented strategy that all stakeholders – from customers to the board – will be able to understand and comprehe...
Apr. 29, 2016 09:30 AM EDT Reads: 951
Increasing IoT connectivity is forcing enterprises to find elegant solutions to organize and visualize all incoming data from these connected devices with re-configurable dashboard widgets to effectively allow rapid decision-making for everything from immediate actions in tactical situations to strategic analysis and reporting. In his session at 18th Cloud Expo, Shikhir Singh, Senior Developer Relations Manager at Sencha, will discuss how to create HTML5 dashboards that interact with IoT devic...
Apr. 29, 2016 09:00 AM EDT Reads: 781
There is an ever-growing explosion of new devices that are connected to the Internet using “cloud” solutions. This rapid growth is creating a massive new demand for efficient access to data. And it’s not just about connecting to that data anymore. This new demand is bringing new issues and challenges and it is important for companies to scale for the coming growth. And with that scaling comes the need for greater security, gathering and data analysis, storage, connectivity and, of course, the...
Apr. 29, 2016 08:45 AM EDT Reads: 552
The increasing popularity of the Internet of Things necessitates that our physical and cognitive relationship with wearable technology will change rapidly in the near future. This advent means logging has become a thing of the past. Before, it was on us to track our own data, but now that data is automatically available. What does this mean for mHealth and the "connected" body? In her session at @ThingsExpo, Lisa Calkins, CEO and co-founder of Amadeus Consulting, will discuss the impact of wea...
Apr. 29, 2016 08:45 AM EDT Reads: 548
Digital payments using wearable devices such as smart watches, fitness trackers, and payment wristbands are an increasing area of focus for industry participants, and consumer acceptance from early trials and deployments has encouraged some of the biggest names in technology and banking to continue their push to drive growth in this nascent market. Wearable payment systems may utilize near field communication (NFC), radio frequency identification (RFID), or quick response (QR) codes and barcodes...
Apr. 29, 2016 08:30 AM EDT Reads: 555
SYS-CON Events announced today that DatacenterDynamics has been named “Media Sponsor” of SYS-CON's 18th International Cloud Expo, which will take place on June 7–9, 2016, at the Javits Center in New York City, NY. DatacenterDynamics is a brand of DCD Group, a global B2B media and publishing company that develops products to help senior professionals in the world's most ICT dependent organizations make risk-based infrastructure and capacity decisions.
Apr. 29, 2016 05:45 AM EDT Reads: 2,399
SYS-CON Events announced today TMCnet has been named “Media Sponsor” of SYS-CON's 18th International Cloud Expo, which will take place on June 7–9, 2016, at the Javits Center in New York City, NY, and the 19th International Cloud Expo, which will take place on November 1–3, 2016, at the Santa Clara Convention Center in Santa Clara, CA. Technology Marketing Corporation (TMC) is the world's leading business-to-business and integrated marketing media company, servicing niche markets within the com...
Apr. 29, 2016 04:45 AM EDT Reads: 2,344
The IoT has the potential to create a renaissance of manufacturing in the US and elsewhere. In his session at 18th Cloud Expo, Florent Solt, CTO and chief architect of Netvibes, will discuss how the expected exponential increase in the amount of data that will be processed, transported, stored, and accessed means there will be a huge demand for smart technologies to deliver it. Florent Solt is the CTO and chief architect of Netvibes. Prior to joining Netvibes in 2007, he co-founded Rift Technol...
Apr. 28, 2016 07:00 PM EDT Reads: 1,475
Join IBM June 8 at 18th Cloud Expo at the Javits Center in New York City, NY, and learn how to innovate like a startup and scale for the enterprise. You need to deliver quality applications faster and cheaper, attract and retain customers with an engaging experience across devices, and seamlessly integrate your enterprise systems. And you can't take 12 months to do it.
Apr. 28, 2016 04:45 PM EDT Reads: 1,737
This is not a small hotel event. It is also not a big vendor party where politicians and entertainers are more important than real content. This is Cloud Expo, the world's longest-running conference and exhibition focused on Cloud Computing and all that it entails. If you want serious presentations and valuable insight about Cloud Computing for three straight days, then register now for Cloud Expo.
Apr. 28, 2016 04:30 PM EDT Reads: 1,628
IoT device adoption is growing at staggering rates, and with it comes opportunity for developers to meet consumer demand for an ever more connected world. Wireless communication is the key part of the encompassing components of any IoT device. Wireless connectivity enhances the device utility at the expense of ease of use and deployment challenges. Since connectivity is fundamental for IoT device development, engineers must understand how to overcome the hurdles inherent in incorporating multipl...
Apr. 28, 2016 02:45 PM EDT Reads: 1,366
Machine Learning helps make complex systems more efficient. By applying advanced Machine Learning techniques such as Cognitive Fingerprinting, wind project operators can utilize these tools to learn from collected data, detect regular patterns, and optimize their own operations. In his session at 18th Cloud Expo, Stuart Gillen, Director of Business Development at SparkCognition, will discuss how research has demonstrated the value of Machine Learning in delivering next generation analytics to im...
Apr. 28, 2016 02:30 PM EDT Reads: 1,564
The paradigm has shifted. A Gartner survey shows that 43% of organizations are using or plan to implement the Internet of Things in 2016. However, not just a handful of companies are still using the old-style ad-hoc trial-and-error ways, unaware of the critical barriers, paint points, traps, and hidden roadblocks. How can you become a winner? In his session at @ThingsExpo, Tony Shan will present a methodical approach to guide the holistic adoption and enablement of IoT implementations. This ov...
Apr. 28, 2016 01:00 PM EDT Reads: 1,510
SYS-CON Events announced today that Stratoscale, the software company developing the next generation data center operating system, will exhibit at SYS-CON's 18th International Cloud Expo®, which will take place on June 7-9, 2016, at the Javits Center in New York City, NY. Stratoscale is revolutionizing the data center with a zero-to-cloud-in-minutes solution. With Stratoscale’s hardware-agnostic, Software Defined Data Center (SDDC) solution to store everything, run anything and scale everywhere...
Apr. 28, 2016 12:45 PM EDT Reads: 1,463
Angular 2 is a complete re-write of the popular framework AngularJS. Programming in Angular 2 is greatly simplified – now it's a component-based well-performing framework. This immersive one-day workshop at 18th Cloud Expo, led by Yakov Fain, a Java Champion and a co-founder of the IT consultancy Farata Systems and the product company SuranceBay, will provide you with everything you wanted to know about Angular 2.
Apr. 28, 2016 12:15 PM EDT Reads: 1,614
SYS-CON Events announced today that Men & Mice, the leading global provider of DNS, DHCP and IP address management overlay solutions, will exhibit at SYS-CON's 18th International Cloud Expo®, which will take place on June 7-9, 2016, at the Javits Center in New York City, NY. The Men & Mice Suite overlay solution is already known for its powerful application in heterogeneous operating environments, enabling enterprises to scale without fuss. Building on a solid range of diverse platform support,...
Apr. 28, 2016 11:30 AM EDT Reads: 2,185
You deployed your app with the Bluemix PaaS and it's gaining some serious traction, so it's time to make some tweaks. Did you design your application in a way that it can scale in the cloud? Were you even thinking about the cloud when you built the app? If not, chances are your app is going to break. Check out this webcast to learn various techniques for designing applications that will scale successfully in Bluemix, for the confidence you need to take your apps to the next level and beyond.
Apr. 28, 2016 11:00 AM EDT Reads: 1,400
SYS-CON Events announced today that Ericsson has been named “Gold Sponsor” of SYS-CON's @ThingsExpo, which will take place on June 7-9, 2016, at the Javits Center in New York, New York. Ericsson is a world leader in the rapidly changing environment of communications technology – providing equipment, software and services to enable transformation through mobility. Some 40 percent of global mobile traffic runs through networks we have supplied. More than 1 billion subscribers around the world re...
Apr. 28, 2016 10:00 AM EDT Reads: 686
So, you bought into the current machine learning craze and went on to collect millions/billions of records from this promising new data source. Now, what do you do with them? Too often, the abundance of data quickly turns into an abundance of problems. How do you extract that "magic essence" from your data without falling into the common pitfalls? In her session at @ThingsExpo, Natalia Ponomareva, Software Engineer at Google, will provide tips on how to be successful in large scale machine lear...
Apr. 28, 2016 10:00 AM EDT Reads: 724