Aerospike: The Database Engine for Real-Time AI and Machine Learning

Table of Contents

1. Introduction
2. The Unique Problem Set of Data in the Age of AI and Machine Learning
3. Aerospike: A Database Engine for Actionable AI and Machine Learning
4. The Challenge of Ingestion: IoT Applications and Aerospike
5. Combining Old and New Technology for Improved Service
6. NGA's Partnership with Amazon Web Services for Text Extraction and Machine Learning
7. Cloudera's Partnership with the Navy for Predictive Maintenance
8. CBP's Use of Data Annotation and Tagging for Better Assessment
9. Pure Storage's Partnership with HHS for Data Integration and Distribution
10. VA's Use of Natural Language Processing and AI for Healthcare
11. Aerospike's Use in Credit Card Processing for Real-Time Decision Making
12. Conclusion

Aerospike: A Database Engine for Actionable AI and Machine Learning

In the age of AI and machine learning, the problem of data has become more complex than ever before. The internet has scaled the issues around data and how to use it effectively in ways that we all knew were coming, but have hit us like a sledgehammer. Most organizations have a lot of data, data that's been collected over the years, and data that's coming in at the moment. Fundamentally, there are lots of different ways You can use data, but you try to extract information from it that can be slowly extracted, or you try to deal with data that's new and deal with it in real-time. The issue is going from the edge back to the Core, and that's where Aerospike comes in.

Aerospike is a database engine that makes a lot of AI and machine learning actionable. It facilitates the issue of going from the edge back to the core. You've got a huge ingestion problem, and if you think about the ingestion problem, one of the IoT applications we're working on deals with an ingestion rate of 12 to 15 terabytes a Second. That's an awful lot of data to sift through, and fundamentally, you can use algorithmic decision engines, and their job is to decide what's the next best action. We're talking about decisions being made in less than an eye Blink, and for people who don't think in terms of milliseconds, an eye blink is 300 milliseconds. So you need to have a data engine that can handle transactions in microseconds, and that's just not where the technology's been. That's what Aerospike can do, from the edge and back to the core.

The Unique Problem Set of Data in the Age of AI and Machine Learning

The problem of data has become more complex than ever before in the age of AI and machine learning. The internet has scaled the issues around data and how to use it effectively in ways that we all knew were coming, but have hit us like a sledgehammer. Most organizations have a lot of data, data that's been collected over the years, and data that's coming in at the moment. Fundamentally, there are lots of different ways you can use data, but you try to extract information from it that can be slowly extracted, or you try to deal with data that's new and deal with it in real-time.

The Challenge of Ingestion: IoT Applications and Aerospike

One of the biggest challenges in dealing with data in the age of AI and machine learning is ingestion. IoT applications, in particular, have a massive ingestion problem, with some applications dealing with an ingestion rate of 12 to 15 terabytes a second. This is where Aerospike comes in. Aerospike is a database engine that can handle transactions in microseconds, making it possible to deal with data in real-time.

Combining Old and New Technology for Improved Service

One of the challenges of dealing with data in the age of AI and machine learning is combining old and new technology to make the whole system fit together. This is particularly important when it comes to improving intimacy with your constituency and providing a better service. Whether it's with the VA or the Department of Defense, it really doesn't make any difference. The goal is to take some of the old technology that's good, combine it with new technology, and make the whole system fit together.

NGA's Partnership with Amazon Web Services for Text Extraction and Machine Learning

The National Geospatial-Intelligence Agency (NGA) has partnered with Amazon Web Services (AWS) to pilot a text extraction machine learning model to start automating the activity of updating aeronautical publications. This is a labor-intensive process that involves reviewing approximately 250,000 pages of information from approximately 200 foreign nation civil aviation programs every month to ensure that there are no updates that would affect the safety of navigation and aeronautical activity for the Department of Defense. The initial pilot was successful, and the NGA is looking at scaling it to multiple other countries, with the potential to save 20,000 hours per year in analytic time and millions of pages of text every year.

Cloudera's Partnership with the Navy for Predictive Maintenance

Cloudera has partnered with the Navy for predictive maintenance for some of their airframes. By leveraging the data from the aircraft, Cloudera was able to predict when certain parts would need to be replaced, reducing maintenance downtime from a month to a week.

CBP's Use of Data Annotation and Tagging for Better Assessment

The U.S. Customs and Border Protection (CBP) is using data annotation and tagging to better assess the data they Collect. By annotating and tagging the data, CBP can look at the metadata on top of the data and build models on top of it to figure out what's happening at different times and assess things better moving forward.

Pure Storage's Partnership with HHS for Data Integration and Distribution

Pure Storage has partnered with the Department of Health and Human Services (HHS) to address the explosion of data growth they've had in regards to biomedical research, patient care, and disease tracking. By creating an on-premises edge Type computing capability and shoring up the on-premises infrastructure of HHS, Pure Storage was able to support the data integration efforts across agencies and enable the application development that drives how well agencies can do their mission.

VA's Use of Natural Language Processing and AI for Healthcare

The Department of Veterans Affairs (VA) is using natural language processing and AI for healthcare. They're using natural language processing to submit and process different forms and collaborating with the Department of Energy on new AI approaches for diseases like cancer, cardiovascular disease, and mental health.

Aerospike's Use in Credit Card Processing for Real-Time Decision Making

Aerospike is being used in credit card processing for real-time decision making. A single business transaction has to go full circle in 750 milliseconds, and the algorithmic decisioning engine has to decide if it's fraud or not in less than half a second. Aerospike can handle transactions in microseconds, making it possible to reduce the incidence of fraud by 30x.

Conclusion

In conclusion, the age of AI and machine learning has brought about unique challenges when it comes to dealing with data. Aerospike is a database engine that makes a lot of AI and machine learning actionable, facilitating the issue of going from the edge back to the core. By combining old and new technology, organizations can improve intimacy with their constituency and provide a better service. Partnerships between government agencies and industry leaders have resulted in successful use cases, such as NGA's partnership with AWS for text extraction and machine learning, Cloudera's partnership with the Navy for predictive maintenance, CBP's use of data annotation and tagging for better assessment, Pure Storage's partnership with HHS for data integration and distribution, and VA's use of natural language processing and AI for healthcare. Aerospike is being used in credit card processing for real-time decision making, reducing the incidence of fraud by 30x.