Transparent, inline
Steer traffic to Hoard-XDN with a bridge, policy route or WCCP-style redirect. No proxy settings on any device — subscribers just get faster downloads.
Transparent HTTP caching · v7.2.54
Hoard-XDN sits inline in your network and quietly hoards the traffic your subscribers pull again and again — OS & game updates, app stores, video, CDN objects — then serves it locally. Your upstream link stops carrying the same bytes twice.
Measured on a live Hoard-XDN deployment. Your numbers depend on your traffic mix.
The problem
Every Windows update, every 80 GB game patch, every viral video and app-store download crosses your upstream link once per subscriber. On a busy access network that is thousands of identical transfers — the exact traffic you are billed for on transit, and the exact traffic that congests your peak hour.
Hoard-XDN breaks that pattern. It keeps one local copy and hands it to everyone else at line rate. The content arrives faster for the subscriber, and it never touches your upstream again.
What's inside
Steer traffic to Hoard-XDN with a bridge, policy route or WCCP-style redirect. No proxy settings on any device — subscribers just get faster downloads.
Hot objects live in a fast memory tier; everything else spills to a disk pool that scales into the tens of terabytes. A 7-day freshness floor keeps cacheable content ready.
An optional Claude-backed advisor tunes caching heuristics to your real traffic. No key? It falls back to fast, deterministic rules — never a hard dependency.
A built-in resolver handles CDN-origin resolution and sinkholing, so cache placement follows the content — not a stale hostname.
Cache hits are DSCP-marked (AF41) so you can prioritise locally-served traffic — savings and speed in one policy.
Simultaneous requests for the same object are collapsed into one origin fetch, and Hoard-XDN serves stale on origin errors — so a flaky upstream never becomes an outage.
Warm the cache with the content you know is coming, so the first subscriber to ask is already a hit.
A live web dashboard shows hit ratio, bytes saved and health in real time, while a self-healing guardian rebuilds the index and watches memory — carrier-grade uptime.
How it works
Install the ISO on a box or VM and point subscriber traffic through it — bridge, policy-route or redirect. Nothing on the client side changes.
Hoard-XDN watches what your network pulls and stores the repeatable objects across its RAM and disk tiers, coalescing duplicate fetches as it goes.
The next request for that object is a HIT — delivered from inside your network at full speed and DSCP-marked so your routers prioritise it.
The dashboard shows hit ratio and bytes offloaded climbing. That's transit bandwidth you no longer buy.
From a live deployment
A single production node. Every network is different — request a quote and we'll estimate your savings from your own traffic profile.
Pricing
No public price list — because a single edge box and a carrier core aren't the same deal. Tell us your scale and we'll quote licensing, and optionally deployment and support.
Single-site ISPs, WISPs & campuses
Growing ISPs & regional providers
Carriers & multi-POP networks
Get a quote
Give us a sense of your scale and we'll come back with a right-sized proposal — licensing, hardware guidance and an estimate of what Hoard-XDN would offload for you.
FAQ
No. Hoard-XDN is transparent. You steer traffic to it at the network level; devices need no proxy settings and users notice nothing but faster downloads.
The repeatable, high-volume traffic that dominates access networks: OS and game updates, app-store and package-repository downloads, software installers, video and CDN-delivered objects. That's exactly the traffic that otherwise crosses your upstream once per user.
It tracks your traffic mix, but repeat-heavy networks cache very well — a live node is running at a 64% request hit ratio and serving ~47% of bytes from cache. We'll estimate your savings from your own profile before you commit.
Hoard-XDN ships as a 2.8 GB installer ISO for bare metal or a VM. It scales with what you give it — a single edge box for a small ISP, up to multi-core carrier nodes with tens of terabytes of cache pool.
Yes — it is vendor-neutral. Cache hits are DSCP-marked (AF41), so any router or firewall can prioritise locally-served traffic with a standard QoS policy. Deployment guidance is part of what we provide.
It's a licensed commercial product from Wizards Services, sold per node with activation. Support tiers run from email up to 24/7 with an SLA, and we can handle the deployment for you. Request a quote to size it.