Resources / The Evolution Blog
5 Advantages of Making Decisions on the Edge
Mike Brooks
What is the Edge?
If we track the history of computing, it started with one large machine. As we progressed to personal computers, for the first time regular users owned the hardware itself. The advent of the cloud led to centralized services specifically from a few providers (Amazon, Microsoft, Google, and IBM). Most companies now rely on the infrastructure, machine learning and compute power from at least one of these providers. The opportunity for growth in the cloud space is now around the Edge. Edge in this context literally means geographical distribution. The computing is done near or at the source of the data. It does not rely on cloud data centers: effectively the cloud is coming to end-users rather than them coming to the cloud. For decision systems, analyzing data at the edge is more cost-effective than moving data from the Edge to another location. Edge use cases mainly centre around IoT / smart devices, but another interesting proposition is using Edge computing to detect fraud, abuse, and security risks.
Content Delivery Networks (CDNs)
As a CDN sits on the Edge between the client (a user-agent, such as your browser or a search engine bot) and the origin server (the server the website resides on), it is in the perfect place to handle rules, machine-learning models, and decisions.
Key-value stores are a common storage solution the CDN providers are adopting. A key-value store is a type of non-relational database that uses a simple key-value method to store and retrieve data at high speeds. We use hashes of details such as device identifier as a lookup mechanism to quickly identify similarities and patterns with previous interactions in the system.
CDNs do make decisions on the edge particularly around bot detection but this raises an important point. If all decisions are made by a CDN, more users are potentially turned away, often based on IP geographical attributes. Many businesses do not want one vendor making all their decisions. They also need multi-dimensional decisions with all available context from how the user interacts downstream. This provides an opportunity for businesses to have another system giving them unparalleled control over their fraud, security, and abuse problems.
Deploying Darwinium on the Edge
The Darwinium Platform brings data science, context and decisioning to the Edge. Deployed via a NGINX plugin or Edge worker, at the CDN level, Darwinium allows businesses to dynamically control their customer journeys from a risk perspective. This can include device and digital data profiling, injecting new content, behavioral biometrics (mouse, touch, and keyboard strokes), IP location, text and images similarity analysis, machine learning and orchestrating third-party scripts and APIs.
Darwinium can be deployed in the following ways:
- NGINX plugin - Deployed/configured directly within your existing NGINX service mesh.
- NGINX Ingress - Provides a managed abstraction for configuring the NGINX plugin.
- Cloudflare worker - Provides decisioning and enrichment at the edge using your Cloudflare CDN.
The Darwinium Platform unites the following 4 capabilities:
Customer Intelligent – Contextualize fraud and abuse intelligence from the packet-to-the-person. Darwinium can inject its own or third-party device profiling JavaScript collecting a huge volume of device signals including behavioral biometrics. Darwinium can also risk-assess content such as images and text. Darwinium image and text similarity allows fuzzy matching for this type of content. This is well suited to abuse use cases – for example looking for malicious images, spam, and other abusive content.
Tailored Journeys – Darwinium is founded on a key concept around profiling user journeys rather than point-in-time interactions. Data mapping directly from the request/response bodies allows Darwinium to dynamically add or remove friction in a customer journey, based on historic and in-session click-stream behavior.
Distributed Orchestration – Bringing data science to the Edge, Darwinium features and models can be built via drag and drop feature editors and integrated notebooks. These features can be propagated across a business’s estate helping multiple business units meaning that Fraud, Security, Abuse, Credit Risk, Marketing and Customer Experience teams can have access to the same single view of the end user.
Decision Control – Models run-when-ready, meaning as soon as the data is available, they execute throughout the customer journey. This allows dynamic strategies based on the business’s risk appetite and executed in real-time.
As mentioned above, one of Darwinium’s key differentiators is around image and text parsing and streaming. Darwinium has built up PDQ hash matching algorithms to allow fuzzy matching on images as well as similar fuzzy matching principles for text. Darwinium has the best of both worlds in this instance. The Darwinium journey editor allows users to specify things from the document, while it is streaming via its real-time engine. The impact is that it allows uniform application of policies across multiple control points, which is a unique approach to image and text matching.
5 Advantages of Making Decisions at the Edge
- Workload - Reduce the number of requests made to the origin server (a reduction in traffic can work more efficiently and improve stability of the server).
- Security - Increase security by filtering out malicious requests/bots. Processing these more locally / at a device level means less likelihood of data breaches from a centralised cloud server.
- Latency – Data processing happens at a more local point, thereby reducing the latency period for the device and improving overall customer experience.
- Costs - Lower cost of data transmission since less data is communicated to the central data warehouse.
- Reliability - The high reliability of connected on-device systems is important in applications that are vulnerable to network disconnections.
Darwinium has chosen to prioritize decisions at the Edge because this approach allows a holistic view of the entire customer journey, meaning the highest level of context is available for every decision. It is also designed to mitigate many of the cons associated with Edge decisioning particularly around data leakage/privacy and device vulnerability. The use cases for Darwinium are extensive and the in-built configuration of the platform allows businesses to tailor capabilities to their specific needs. CDNs represent a useful deployment option beyond NGINX. Deployment via Cloudflare is available now, with other CDN integrations already on the roadmap.