Open Structure Planning - Part One

This is part one of the Series on Management and Proper contingency planning of your Websites.

Open Structure consists of you knowing where everything are, what they do, by the subdomains and such, and also main importantly, what to do in an emergency.

Emergencies can be classed as :
Deliberate Denial Of Service or DDOS Attacks (aka Hackers Bombardment)
Suspension of Domain Services
Server / Service Provider Outage
Natural Disasters
Domain Hostile Take Over.

DDOS are normal, and like an example is Comment Spamming. A Good process is to remove the naming sequence for WP-COMMENTS-POST in your INDEX or PAGE.PHP to something else. It’s located in the form entity.

Suspension of Services can occur when a rouge script, hacked or unintentional overloads the resources and takes over more than 30 seconds of Processor Time in each instance. This is a norm in any Hosting Environment. The Limit may also be 30% of the SINGLE CPU LOAD at any one time. If you utilize FEEDS or XML Calls, a delay in your uplink feed causes this issue predominantly. So, do a remote call as a CRON JOB instead.

Sometimes, it may be also our own fault, like internal ads showing some sort of Pornography. This will lead to suspension. Another mistake is PayPal auto payment system failed to debit the funds from our account or credit card due to Non Sufficient Funds (NSF) or Credit Card companies not authorising them for some stupid reason what so ever.

Server Outage can occur as this relies on Network Attached Storage devices failing due to Corruption on the Data heads of Hard Drive Failures, as it will break down even a Raid-10 system. Powweb had this issue last year, and killed the entire network , even with Load Balancing in place, as a good load balancer,only distributes the load thru within the cluster of servers but they still dependent on the NAS, and as we all know it, NAS uses Hard Drives, and they are made of spinning moving discs and motors and rotors. To get a Raid rebooted or remounted back can take HOURS, depending on how much damage was the corruption, and with the provision of a proper snapshot of the data embedded within. What is made worse, that the Database Servers also rely on this, and any non-async of the data will kill the clusters.

Natural Disasters like Flood and Earthquake causes alot of damage. In California alone, a single shock can cost the population over $20Billion in damages. Even with a shock proof building and power generators backed up, the Power Grid and Telephone lines are in the ground, and a knockout at the Exchange will kill the feeds. It’s no good to hear that an ISP has 30 or 100 connectivity to the outside world when a 5.0 on the Richter Scale will kill the city Telecommunications Backbone.

Even a knock on the Far Eastern Plates, like what happened to Taiwan recently killed ALL the UPLINK UNDERSEA CABLES at that point of time. We know that most of the economy now depends on Asia , and without the Net Access to our clientele there, we are roasted!~ Some ISPs are better equipped with redundant systems and routes, and a tertiary communications layer is enabled. Though its very technical, but it’s quite simple to implement, but costly, at the rate of about $10,000 a minute.

Hostile Domain Takeover is when your nameservers are changed and sites removed , due to either NON PAYMENT of RENEWAL FEES or simply, being taken over.

Enough Nightmares already??