Scaling with Fully Integrated Node.js Applications

Fully Integrated

Harper delivers the power of a database, web cache, messaging, and application functions in a single, convenient Node.js package. This allows your web applications to be delivered quickly and maintained easily into the future. But can Node.js support large scale? To answer that with a resounding ‘yes’, let’s explore how HarperDB easily scales with large sites and apps.

Case Studies

HarperDB Box with Node.js box coming out of it

Fused-Stack Application

This video showcases a commerce application with Next.js hosting, backend data, and third-party AI personalization, all delivered with Harper. It highlights the impressive loading speeds, with LCP times below 0.2 seconds, emphasizing the dynamic data pulled from the Harper database. Additionally, it present a unique feature of product personalization using OpenAI, where product descriptions are tailored based on user attributes.

Test Demo Site

Tailored Experiences

In today's world, users expect sites and applications tailored to their needs – although this may sound obvious, it is not a trivial problem. When I log in to shop, I'd like to see my shopping cart and get recommendations based on my prior buying behavior. When I track my runs online, I'd like to see my own and perhaps those of my friends. Customizing the application to the data that pertains to me is the name of the game.

Faster Queries, More Revenue

Customization means session-specific queries - Queries for data take time, and users that wait for their shopping cart are users that bounce to the competitor. Speed is essential, and Harper achieves unmatched performance with a few very cunning tricks.

Performance Considerations

The time it takes for the user to make a request and the data appearing on the site – is a complex interplay of application delivery characteristics, distance to the user, and the volume of data needed to deliver the custom experience.

Like a CDN but for Dynamic Data

Edge computing allows this customization work to be done closer to the user, much like traditional Content Delivery Networks brought static content closer to the user to reduce this distance factor. Harper is doing this same but for dynamic data.

Query Acceleration

Since data can be expensive to query, Harper makes use of a very high-performance C++ storage engine (LMDB) that employs smart in-memory caching to ensure that customization queries run fast, return data quickly, and answers can be modified for the user right on the same server node giving the user the customized experience they have come to expect.

Horizontal Scale

These optimizations pay off: individual instances support 120,000 reads per second and up to 20,000 writes, allowing clusters of only a handful of Harper instances to support millions of concurrent users.

Cut Overhead, Slash Latency

Harper is unique in that it combines messaging, memory caching, application delivery, and database all in one package. This smartly avoids the overhead that is typically associated with federating each of these tasks to separate components. Ensuring minimal latency and the best user experience.

Simple Scale

By collapsing the technology stack into a single application delivery platform, scale is no longer an issue.

Smarter Scale

By removing this overhead, smaller instances can be used instead of costly ‘high-memory’ instances.

Distributed Scale

Deploy smaller horizontally scaled instances that are geographically distributed and closer to the end-users.

Dev Center

Our Story

Build

Faster Applications