# I spent the week mapping web architectures

Canonical: https://brew.new/browse/templates/email/pt1_k97gqe55myecey7dzda8a0ypn18e2hgg

Brand: harperdb.io
Category: newsletter

![Preview of I spent the week mapping web architectures](https://cdn.brew.new/email-preview-70c6dcc67f5aee0d-tracking_r57xmwfwdshzm68p6712pvfec58dtp71-1788998236527.png)

## Email content

Hi there,

I spent last week going deep on architectures for web personalization - static, edge, server, client, and unified - trying to pin down which is best.

Performance came down to two distances: how far the user is from the compute and how far the compute is from the data. Most architectures win on one and lose on the other. Only one keeps both close.

You can see the architectures and their latency-inducing network hops visualized side by side here:

0-Comparison-1

See where each architecture lands → Here

What drove it home for me was the benchmark. When we tested Harper against the full Vercel stack on live, personalized reads, Harper came out up to 14x faster - same app, same data, just less distance to cross.

In case you missed it last week, the numbers are all here → here

Aleks @ Harper

Harper, 2420 17th Street, Suite 270, Denver, Colorado 80202, United States

Unsubscribe

Manage preferences

[Open and remix this design](https://brew.new/browse/templates/email/pt1_k97gqe55myecey7dzda8a0ypn18e2hgg)

[Browse email designs](https://brew.new/browse/templates)