AXPEED DB in 60 Minutes

1  What AXPEED DB Is

AXPEED DB is a multi-paradigm database built from the ground up by TNZ Soft. One binary. One server. Eight ways to use it.

And it speaks seven wire protocols — PostgreSQL, MySQL, Redis, MongoDB, Oracle TNS, SQL Server TDS, and REST — so your existing applications connect without changing their database driver.

2  Install and Connect

# Linux (Debian/Ubuntu)
sudo dpkg -i axpeed-db_1.4.0h_amd64.deb
sudo systemctl start tnzsoldb

# Connect with the SQL shell
tnzsql -h localhost -p 15432 -U admin -d soldb

tnzsql (1.4.0h)
Type "help" for help.

soldb=#

That is it. The database is running. The default port is 15432, the default database is soldb, and the SQL shell is tnzsql.

3  Your First Queries

-- Create a table
CREATE TABLE products (
    id       SERIAL PRIMARY KEY,
    name     TEXT NOT NULL,
    price    NUMERIC(10,2),
    category TEXT,
    specs    JSONB            -- JSON data stored alongside SQL columns
);

-- Insert some products
INSERT INTO products (name, price, category, specs) VALUES
('Laptop Pro 15',   189000, 'Electronics', '{"ram":"16GB","storage":"512GB SSD","brand":"TechCo"}'),
('Wireless Mouse',    2500, 'Electronics', '{"brand":"ClickMax","dpi":3200}'),
('Standing Desk',    45000, 'Furniture',   '{"height":"72-120cm","material":"bamboo"}'),
('Ergonomic Chair',  38000, 'Furniture',   '{"weight_limit":"150kg","adjustable":true}'),
('4K Monitor',       62000, 'Electronics', '{"size":"27 inch","panel":"IPS","refresh":"144Hz"}');

-- Standard SQL — works exactly as you expect
SELECT name, price FROM products WHERE price > 40000 ORDER BY price DESC;

 name           |  price
----------------+---------
 Laptop Pro 15  | 189000
 4K Monitor     |  62000
 Standing Desk  |  45000
(3 rows)

-- Aggregate with GROUP BY
SELECT category, COUNT(*) AS items, AVG(price)::NUMERIC(10,0) AS avg_price
FROM products GROUP BY category;

 category    | items | avg_price
-------------+-------+----------
 Electronics |     3 |    84500
 Furniture   |     2 |    41500
(2 rows)

Query JSON Inside SQL

No separate document database needed. AXPEED stores JSONB in binary format with GIN indexing — query it with SQL operators:

-- Find all products by brand "TechCo"
SELECT name, price, specs->>'brand' AS brand
FROM products
WHERE specs->>'brand' = 'TechCo';

 name          |  price  | brand
---------------+---------+--------
 Laptop Pro 15 | 189000  | TechCo
(1 row)

-- Find electronics with specific JSON attributes
SELECT name, specs->>'size' AS screen, specs->>'refresh' AS refresh
FROM products
WHERE category = 'Electronics' AND specs ? 'size';

 name       | screen  | refresh
------------+---------+---------
 4K Monitor | 27 inch | 144Hz
(1 row)

4  The Performance Numbers

These are independently verified benchmarks — AXPEED DB v1.4.0h versus PostgreSQL 16.15, same hardware (Intel Xeon 2.80 GHz, 2 cores, 7.8 GB RAM), same queries, same data:

And on a single core with 10 million records:

5  In-Database Machine Learning — 470× Faster

This is the number that changes the conversation. When an application needs an ML prediction from data in a database, the traditional approach is:

99.8% of the time is moving data. Only 0.2% is actual machine learning.

AXPEED runs the ML function inside the engine, next to the data. No serialization. No transfer. No conversion:

-- Train a model (one time)
CREATE MODEL price_predictor
USING linear_regression
ON (SELECT sqft, bedrooms, bathrooms, age, price FROM houses);

-- Predict — one line of SQL, 27 milliseconds
SELECT address, predict_price(sqft, bedrooms, bathrooms, age) AS predicted
FROM new_listings
ORDER BY predicted DESC;

 address                  | predicted
--------------------------+----------
 45 Clifton Block 5       | 32500000
 12 DHA Phase 6           | 28100000
 Plot 7, Gulshan-e-Iqbal  | 15800000
(3 rows)

6  Security That Does Not Exist Elsewhere

-- Enable Field DNA on a table
CREATE DNA TABLE financial_records (
    txn_id    SERIAL PRIMARY KEY,
    account   TEXT NOT NULL,
    amount    NUMERIC(12,2),
    recorded  TIMESTAMP DEFAULT NOW()
);

INSERT INTO financial_records (account, amount) VALUES ('ACC-1001', 50000);

-- Verify nobody has tampered with the data
SELECT *, VERIFY_DNA(*) AS integrity FROM financial_records;

 txn_id | account  | amount   | recorded            | integrity
--------+----------+----------+---------------------+----------
      1 | ACC-1001 | 50000.00 | 2026-10-24 09:15:03 | VALID
(1 row)

If anyone changes even one digit of amount — even a superuser — VERIFY_DNA(*) returns TAMPERED. No other database provides this at the individual field level.

7  Vector Search — AI-Native

-- Add embeddings to products (from any AI model — OpenAI, Cohere, etc.)
ALTER TABLE products ADD COLUMN embedding VECTOR(3);

UPDATE products SET embedding = '[0.82, 0.15, 0.91]' WHERE name = 'Laptop Pro 15';
UPDATE products SET embedding = '[0.79, 0.18, 0.88]' WHERE name = '4K Monitor';
UPDATE products SET embedding = '[0.12, 0.95, 0.08]' WHERE name = 'Standing Desk';
UPDATE products SET embedding = '[0.15, 0.92, 0.11]' WHERE name = 'Ergonomic Chair';
UPDATE products SET embedding = '[0.75, 0.22, 0.85]' WHERE name = 'Wireless Mouse';

-- "Show me products similar to the Laptop"
SELECT name, price, embedding <=> '[0.82, 0.15, 0.91]' AS distance
FROM products
WHERE name != 'Laptop Pro 15'
ORDER BY distance
LIMIT 3;

 name           | price  | distance
----------------+--------+----------
 4K Monitor     | 62000  | 0.0014
 Wireless Mouse |  2500  | 0.0038
 Ergonomic Chair| 38000  | 0.9412
(3 rows)

The monitor and mouse are close in embedding space — they are similar products. The chair is far away. This is how recommendation engines work. No separate vector database. No data pipeline. One SQL query.

8  Seven Protocols, One Server

Your MySQL application connects to AXPEED on port 3306 and does not know it is talking to a different database. Your Redis client reads from AXPEED's cache. Your REST frontend calls the database directly with HTTP. One server replaces five.

# PostgreSQL client (port 15432)
psql -h axpeed-server -p 15432 -U admin -d soldb

# MySQL client (port 3306)
mysql -h axpeed-server -P 3306 -u admin -p soldb

# Redis client
redis-cli -h axpeed-server

# REST API (any language, any HTTP client)
curl http://axpeed-server:8081/api/query \
  -d '{"sql":"SELECT * FROM products WHERE price > 40000"}'

9  Licensing That Does Not Punish Growth

Most commercial databases charge per processor core. Upgrade your server from 16 cores to 64 cores — your license cost quadruples, even if your application still uses 30 connections.

AXPEED charges per connection. Upgrade hardware whenever you want. The license cost does not change.

The OPEN tier is free with no expiration and no feature restrictions. Every student, startup, and developer has access to the full engine. If your application outgrows 5 connections, upgrade. Your data stays exactly where it is.

10  What the Full Book Teaches You

This preview showed you 60 minutes of content. The full book has 34 chapters across 8 volumes — a complete database education from first query to production deployment:

Every chapter includes: prerequisite review, learning goals, step-by-step examples you run yourself, real-world scenarios, career builder sections with resume lines and interview questions, exercises with full solutions, and problems rated by difficulty.

Running project: Across 25 build steps, you construct SolPoint — a complete AI-powered point-of-sale system with relational integrity, JSONB product attributes, full-text search, vector similarity recommendations, demand prediction, stockout alerts, neural network product classification, Field DNA audit trails, backup, replication, and security. One database. Zero external services.