Algo-Logic's Order Book for Ad Tech
Today’s Advertising Technology (Ad Tech) requires time to parse a bid request, look up relevant data about a user, evaluate an opportunity, decide on the value, and then bid a price to serve an impression. Due to the overhead of Real Time Bidding (RTB), advertisers only get one chance to bid their Cost Per Impression (CPI). The RTB protocol serves hundreds of billions of ads daily on about 100 demand-side platforms (DSPs), moving hundreds of Terabytes (TB) of data back and forth through ad exchanges to serve ad impressions.
At the OpenPOWER Summit 2016, John Lockwood, CEO, presented a session on how to accelerate RTB in the Ad Tech industry. Tremendous improvements can be achieved by applying Gateware Defined Networking® (GDN) technology to these issues in a similar way as it is applied to financial markets. Technologies already used in financial markets provide excellent transparency and price discovery that can be applied to RTB. By monitoring live market data feeds and constructing an order book with Field Programmable Gate Array (FPGA) logic, multiple auctions can be performed with sub-microsecond latency.
Algo-Logic’s Order Book can be applied to Real-Time Bidding for a new and improved Ad Tech exchange. The Order Book efficiently tracks all open orders and offloads sorting to FPGA hardware in order to sustain high volumes of traffic. As a result, the market is fair, the exchange provides ample liquidity, and transactions can be profitable for all participants.
View John's session at the OpenPOWER Summit on "Moving Ad Exchanges from Main Street to Wall Street."
Algo-Logic's Key-Value Store for Ad Tech
The Ad Tech industry also collects an enormous amount of data, and Algo-Logic's Key-Value Store (KVS) provides a simple and scalable means to store and retrieve distributed data. In datacenters, high performance KVS tables allow hundreds or thousands of machines to easily share data by simply associating values with keys and allowing client machines to read and write those keys and values over standard high-speed Ethernet.
Data Management Platforms (DMPs) combine the online and offline data to create digital audience segments used in programmatic bidding. Some examples of user targeting variables include: User ID, Cookie ID, IP Address, Device ID, In Market Segment (i.e. auto, travel, mortgage), and income range.
Recently at the Intel Developer Forum 2016 (IDF) in San Francisco, Algo-Logic presented the capabilities of the Key-Value Store running on an Intel® Xeon® + FPGA server fed by data collected by our Black Diamond (a robust, real-time measurement and monitoring system) in a sensor fusion demo. We observed how machine-to-machine real-time data can be efficiently collected, analyzed, and shared between network-attached devices in the cloud with sub-microsecond latency for industrial use-cases such as RTB for Ad Tech, trading, asset monitoring, telecommunications, and NoSQL database acceleration. Check out the topology of the Black Diamond feeding the Key-Value Store, and the resulting candlestick charts. By accelerating cloud infrastructure with programmable logic, network operators can save significant time and compute energy while increasing the throughput of data processing.