Session 1: Data, Data Everywhere, and Lots of Drops to Drink
- Data coming from everywhere, but it is not necessarily useful in its current form. Must be massaged to be as beneficial as possible.
- Use case: 3 large apartment buildings in emerging country. They want to address how residents can use their electrical, heating, and cooling systems efficiently.
- Key is to collect this historical data to feed it back into the analytics system and the user.
- Graphed data starts to show a story. Are the households working? How many individuals live there? What are their habits based on energy usage graphs (washing machine use, etc.)?
- Multiple building blocks of technology to build this system. Technology alone is not sufficient.
- Incoming data is turned into objects via Message Hub and cleaned. Tags and metadata attached.
- After the data is massaged and cleaned, data filtering is done at the object and string level.
- Data is then feed into Cognitive Analytics for analysis.
Session 2: Optimize Your Business Cost-Performance with Composable Systems
- CPUs are starting to get in the way. They are not the most efficient way to get mass amounts of data processed. Accelerators are taking on more and more of the load.
- The faster the data gets through the CPU and to the accelerators, the better the performance.
- Accelerators then talk to each other instead of talking with or going through the CPU.
- Composable systems are software-defined systems.
- Uses REST APIs to allocate resources as needed.
- Performance is bad in traditional systems because of the layers that the data has to travel through. Reliability is worse due to more components in use.
- Swapping out hardware and using different types is much easier in these scenarios. This is especially important due to differing upgrade cycles on hardware.
Session 3: Innovations and Research Challenges in IoT
- Interactions are much more important today. People want to interact with systems like they do with other people. Expectations are much higher.
- Conversations need to be stateful. Remembers what happened previously or where one left off.
- AR will be playing a important role as well. Showed an example of QR codes on industrial machines with AR overlay when looking at it.
- Data doesn’t have to reside in the cloud to process it. Compute power is rising at the edge and can harness both cloud and on-prem.
- Security is important when processing IoT data. Some of the risks can be midigated by only sending filtered data from the device to the cloud and not all of it.
- Block chain can be used in these cases as well. It doesn’t just have to be used for financial transactions.
Session 4: Enabling End-to-End Accessiblity Automation and DevOps in the Cloud Era
- Providing accessibility for all improves the workplace. Not just for the individual or individuals with the disabilities. Improved communication, collaboration, and efficiency.
- Utilize users’ thoughts, experiences, and feedback.
- Automation of accessibility tests are critical throughout the testing and implementation phases. Extensions have been designed for Chrome, Firefox, etc. to help in these phases.
Session 5: API Lifecycle Innovations
- APIs are essential for today’s cloud platforms. It enables fast and efficient ways to market for developers.
Session 6: The Serverless Revolution Continues: OpenWhisk
- Serverless is attractive for the same reasons virtualization and containers are desirable.
- Multiple languages are supported.
- Docker containers can be utilized as well.
- Triggers can be mapped to actions based on rules.
- These sets are then brought together into packages (collections of Actions and Feeds)
- Examples of this include an action that converts weather payloads into different languages and spits it out to HTML or JSON.
- Integration of OpenWhisk, Watson Conversations, and Bluemix can connect right into Slack.
Session 7: Innovations in End-to-End Application Development in Swift: Demo
- There are credential, session, websocket, and many other pluggable sets of code.
- Looking at adding support for DB2 and MongoDB.
Session 8: Cognitive Security Services for a Cloud Platform
- Cognitive security brings about the idea that we will look to move away from alerts and to actions directly carried out by a machine. Alerts should only be necessary when a machine is unable to handle a situation.
- This will not be primarily machine only or human only, it is a harmony of the two.
- The data in this space is too large when millions of hosts are generating TBs of data per day and even faster in certain cases.
- There is also a tremendous amount of data that is created for human digestion like analyst reporting and white papers.
- A model is being built to designate usage only to hardware that has these capabilities.
Session 9: Cognitive IoT Applications at the Edge
Session 10: The Serverless Revolution Continues: The Latest From OpenWhisk
- OpenWhisk is a platform that allows a developer to code without setting up an infrastructure. It was built in BlueMix.
- IBM Watson Conversations allows you to give examples of the intent of a conversation in order to best learn what is desired.
- The code used here is able to be used for voice, text, etc. It does not have to be rebuilt to handle it.
Session 11: Software-Defined Power for Cloud Datacenters
- Power capping can be implemented on lower priority servers to move the power needs to the higher priority servers.
Session 12: Demo of Cognitive Security Services for a Cloud Platform
Session 13: Person-Centric IoT
- IBM has partnered with Harmon Internal and BMW to supply voice interaction in cars and audio systems. Examples include hotel automation in rooms with voice commands for drapes, temperature, and lights. Gesture recognition with smart watches as well.
- Examples include extreme conditions, such as heat exposure, employee tracking via beacons in warehouses, proximity alerts when driving forklifts, and fatigue alerts.