Streaming Data For Ad Tech | VoltDB
page-template-default,page,page-id-6356,page-child,parent-pageid-12986,mkd-core-1.0,highrise-ver-1.0,,mkd-smooth-page-transitions,mkd-ajax,mkd-grid-1300,mkd-blog-installed,mkd-header-standard,mkd-sticky-header-on-scroll-up,mkd-default-mobile-header,mkd-sticky-up-mobile-header,mkd-dropdown-slide-from-bottom,mkd-dark-header,mkd-header-style-on-scroll,mkd-full-width-wide-menu,mkd-header-standard-in-grid-shadow-disable,mkd-search-dropdown,wpb-js-composer js-comp-ver-6.1,vc_responsive

Ad Tech

VoltDB / Solutions / Ad Tech

Ad Tech

Programmatic Advertising Technology (Ad Tech) is all about speed, scale, accuracy and transparency. Companies in this space rely on being able to process high volume transactional workloads with  millisecond response times. When you visit a website and notice an ad being rendered on the page header top or embedded natively on the page, you are seeing a seamless orchestration of data between Publishers, DSPs, SSPs, Ad exchanges, DMPs and Agency trading desks. With the increase in demand for targeted display advertising, advertisers are looking for reassurances they are getting the right value for their advertising spend and that their programmatic suppliers are not taking too big a cut of their budgets — money that could be better spent elsewhere.

The VoltDB Fast Data platform has not only reduced our footprint, but provided us with the ability to meet the real-time demands of native programmatic advertising while ensuring the accuracy of a growing number of transactions to keep our clients’ ad budgets and actual spend aligned. — Michael Harroun, Head of Backend Engineering, TripleLift

Providers of programmatic platforms manage display advertising for the world’s leading brands. These platforms enable accurate matchmaking between bidders and publishers, maximizing campaign effectiveness.

The Challenge

Programmatic advertising providers must deliver ads at massive scale and with precise targeting capabilities while ensuring accurate campaign balance tracking. The ability to process bid requests at scale and with high precision requires the ability to correlate campaign, budget and user data in a meaningful manner. The complexity of the process and the constraints on latency SLAs lead to the following challenges:

  1. Effective campaign management
  2. Maintaining high-precision targeting
  3. Lack of transparency in true winning bids
  4. Ineffective matchmaking leading to fall-back bids
  5. Accurate globally distributed spend management.

Data architecture choices that create these challenges are typically:

  1. NoSQL databases, which make data transparency difficult or impossible
  2. Multi-tier data processing, which significantly increases latency, unnecessarily
  3. Eventual consistency models, which lead to inaccuracies in spend management and campaign performance.

The Solution

Managing programmatic bids and auctions at scale requires the ability to process and act on events and data at the rate at which they are generated. VoltDB meets the demands placed on data processing in a predictably scalable manner while ensuring latency SLAs are not violated. VoltDB reduces the complexity of data architectures by consolidating processing logic and data storage into a single layer, reducing latency. Global AdTech organizations benefit from VoltDB’s Multi-Site Cross-Datacenter Replication (XDCR) for managing global ad campaign spend.

A content delivery network (CDN) provides content from advertisers, ad networks, publishers, hosting providers, gaming companies, and mobile providers to end users. The main premise is to make websites faster by localizing content and delivering it faster on any device, no matter the file size.

The Challenge

CDNs bill customers based on website content “views.” The system implemented must provide absolute precision in counting content views at high speed and scale so customers are billed appropriately. A system that does not meet the speed and scale requirements will result in billing issues. Under-billing means losing money; over-billing creates customer satisfaction problems.

The Solution

VoltDB processes streaming data in real-time while maintaining the “exactly once” semantic, which ensures data is not over- or under-counted to deliver accurate billing.

Another option is building a stack of Apache projects, such as ZooKeeper, Kafka, Storm (or Spark) and Cassandra, or the Lambda Architecture, to solve fast data challenges. However, these approaches require a massive development effort to patch together a solution, sacrificing performance, ACID guarantees, and ease of implementation.

Bottom Line

VoltDB helps content delivery networks concentrate development resources on building applications rather than on the data processing infrastructure. It is purpose-built for companies that need a fast, scalable database capable of processing streaming data and making in-transaction decisions with sub-millisecond latency. VoltDB provides a simpler design: 1/10th the complexity of the Apache stack; precise results with exactly-once semantics; and a hardware-efficient solution.