Mark Van de Wiel Abstract dots multi-colored dots on black background

The Case for Real-Time Analytics

Imagine you are driving your family to a holiday destination. You are looking forward to a relaxing couple of weeks at the beach. Four hours into the journey, as you get close to your destination, the dashboard indicates a mechanical issue. But you have not yet noticed a difference in your car’s driving.

What will you do?

A car is a complex product. It is built to be resilient. It is built to protect its passengers. Warning signals are designed to pro-actively indicate a possible future problem. What if the warning is due to a faulty sensor? You perform a risk analysis as you consider a few obvious options on what to do next:

  • You could stop the car and call for repair services. But you consider the delay in the journey and the current weather conditions, as well as the amount of time for help to arrive.
  • You may decide that you will find a local repair shop specializing in your car’s brand during your vacation.
  • You could ignore the warning for now and plan to take the car to the repair shop after your vacation.
  • Or you ignore the warning altogether and keep driving until the next scheduled service.

There should be a better way…Real-Time Analytics

Real-time analytics is a widely used IT term that can mean different things to different people. To some, it means extremely fast query response times, even if the data being queried resides in a data warehouse that hasn’t been updated in days. To others, it can mean embedding a monthly BI report into a workflow so that the user has relevant and sufficiently current information needed to complete a business process.

When we talk about real-time analytics in the context of how HVR customers are using it, we apply a more literal meaning. Our customers use HVR data replication software to continuously analyze changing data generated by transactional systems, machines, sensors, mobile devices, and websites. The latency of the data is typically measured in seconds.

For example, Lufthansa uses HVR to deliver real-time flight planning services to national and international airlines worldwide. Airline operations teams and pilots receive up-to-the-second information about routes, flight paths, air traffic, flight restrictions, weather conditions, and more so that they can adjust schedules, routes, and flight patterns in real-time.

Similar applications can be found in telecommunications, healthcare, law enforcement, public safety, energy and utilities, financial services, and other industries. New applications are primarily driven by big data innovations, the Internet of things, and mobile computing.

Different Approaches to Implementing a Real-time Analytics Solution

There are several different approaches to implementing a real-time analytics solution. The most common include query federation, event and trigger-based replication, and log-based change data capture (CDC).

Our experience, working with customers around the world spanning a broad cross-section of industries, has proven that for the vast majority of use cases, log-based CDC is by far the most practical, efficient, and cost-effective approach. It is transactional, reliable, minimizes network traffic, and non-invasive to source and target systems.

Predictive Analytics

Predictive analytics “encompasses a variety of statistical techniques from data mining, predictive modeling, and machine learning, that analyze current and historical facts to make predictions about future or otherwise unknown events.”

With real-time use cases, there are examples of predictive analytics use cases across industries. Insurance companies use predictive analytics to calculate premiums so that they will be able to fulfill their future obligations. An online retailer uses predictive analytics on its infrastructure to ensure the website remains responsive while running promotions. SaaS providers have to ensure they can deliver their service reliably 24×7. A bank wants to ensure the ATM has enough cash until the next scheduled refill.

In our daily lives, we also, often implicitly, use predictive analysis. Our often not-so-scientific approach is typically called “gut feel.” For example, you want to ensure you buy enough fresh fruits until your next planned run to the grocery store. Or you assess the likelihood you will endanger your family or introduce long-term car damage when you continue driving your car despite a warning indicator.

Real-Time Predictive Analytics

As organizations go through their digital transformations, they recognize the value to perform predictive analytics closer and closer to real-time. The classic example is the call center use case: how likely is a customer to churn based on the information (s)he provides. Again there are examples across industries.

  • Hospitals want to understand a patient’s illness based on evolving symptoms. Doctors want to predict what side effects may result from applying certain medications, given a patient’s profile.
  • Stock market traders continuously try to minimize the risk of exposure given an influx of information.
  • A grocery store will try to optimize the supply chain and inventory on perishable products to avoid running out of a product and avoid losses due to products going bad.

Going back to the example of our vacation: if only real-time predictive analysis were an option. Let’s say the warning signal is for a brake pad wearing down.

  • Did the brake pad lose half its thickness in the last 30 minutes or over the previous three years?
  • What does the manufacturer know about brake pad wear on cars of the same model?
  • Have wear and tear issues been reported with brake pads of a similar kind?

Your gut feels assessment will have to guess the answers to these and other questions related to the warning signal. Likewise, you may have to think about the impact of the loss of brake power on one out of four wheels. This leads to stress – quite the opposite of the relaxing vacation you were looking forward to.

May the digital transformation continue…

If you are interested in learning more about how organizations are applying real-time analytics, the pros and cons of various architectural approaches, and how HVR works in different scenarios, check out our solution brief on the topic.

For more general inquiries, feel free to contact us.

About Mark

Mark Van de Wiel is the CTO for HVR. He has a strong background in data replication as well as real-time Business Intelligence and analytics.

Test drive
Contact us