Uncategorized

Oracle Mysql Cluster 7.4 features

I’ve explored the features of Mysql Cluster 7.2 and 7.3 and now Oracle is coming out with 7.4 with very nice improvements.

This is yet another reason to go for baremetal (softlayer or others) and setup 4 node machine (or more) and enjoy the features of fairly priced or free mysql cluster.

These are the features listed from a Webinar :

 

  • Performance enhancements: MySQL Cluster 7.4 delivers a 50 percent performance improvement over MySQL Cluster 7.3 for Read-Only workloads and 40 percent improvement for Read/Write operations. Performance improvements are available through SQL or any of the native NoSQL APIs supported by MySQL Cluster. New records of 200 Million NoSQL Queries Per Second (QPS) and 2.5 Million SQL QPS were established.
  • New geographic redundancy features for high availability across data centers: The latest release provides rollback of any conflicting transactions – enabling full active-active, update-anywhere replication between geographically distant clusters, with applications able to send reads and writes to any site without compromising on consistency.
  • Faster online maintenance and recovery operationsincluding the ability to roll out software upgrades as much as 5x faster.
  • Enhanced distributed monitoring of memory usage and database operations.

Share This:

architecture, coding

Open sourcing my lottery server

2007 project involving Message Driven Beans (EJB) running on Enterprise Java container (Jboss or Glassfish) where games are processed and “played” as they come in through the Queue allowing scalability and consistency.

 

Source code for both game logic and EJB server can be found on github:

Github to Lottery Server

Share This:

javascript

Add async to script tag and enjoy 10% performance in page load

Chrome 41+ supports

 

<script async src="app.min.js">

 

Which allows script(s) to be parsed in parallel to other scripts therefore end user will experience slightly faster page load.

Share This:

bigdata

Creating Real Time topology in Apache Storm is amazing

Hiding away all the complexities of running nodes (hidden by zookeeper) and providing parallelism for Bolts (units of work in Storm) , as well as creating new Spouts (units of data steam, e.g. redis data, twitter stream or anything you want) allows you to focus on your real time product.

Once the bolt processes a string , adding a counter to it and pushing it to redis topic (using pub/sub).

On the other end, Web app is using d3 to grab all the content coming from a stream (simple python client connected to redis emits data)

 

That’s amazing.

 

 

Share This:

coding, mongodb, nodejs, Uncategorized

Recently open sourced project – toronto open data

The motivation behind this project is

 

  • Building a very capable api with geo lookups
  • Toronto Open Data api -specifically city services is very slow
  • User interface as a proof of concept which exploits the use of new API
  • Open sourced for others to use and improve

Technology stack:

 

  • MongoDB
  • NodeJS, rest api, mongoose
  • AngularJS (bootstrap, routes)

 

GiHub link

Share This:

Uncategorized

Async javascript lib for node and browsers has a better version

https://github.com/suguru03/neo-async

 

Great replacement for current users of async.js which hasn’t been updated for a while.

In a large pharma oriented app, i’ve noticed speedup where async.parallel happens.

 

 

Share This: