CI Tools and Best Practices in the Cloud

Continuous Integration

Subscribe to Continuous Integration: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get Continuous Integration: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Continuous Integration Authors: Yeshim Deniz, Elizabeth White, Pat Romanski, Liz McMillan, Mehdi Daoudi

Article

Moving Smoothly Through the Post-Disaster Panic

Remaining calm when systems go down

When asked how they protect their corporate data, most business owners will tell you that they back up on - at minimum - a daily schedule.

But is this really enough? Well... it all depends on WHY you're backing up the data in the first place.

The "why" of data protection is one of those obscure, poorly understood concepts that suddenly become very important during an emergency. That's why you need to frequently test your backups and your recovery plans using various disaster scenarios.

If your servers were to crash right now, what effect would this have on your stakeholders, customers, partners and employees?

  • Would a few hours of downtime tempt your customers to test out a competitor's offering?
  • Would a few hours of downtime cause employees to miss important deadlines?
  • Would a few hours of downtime and data loss lead to fines or contract breaches due to late shipments or lost transactions?
  • How will you respond to bad press or social media resulting from the outage or accident?
  • Before you can load your old backups, you must first re-build and configure the server from scratch. How long will that take?

These are just a few of the most important questions that need to be considered as part of your overall data disaster plan. In an emergency, everyone in the company should know their roles, and instructions should be clearly followed by all involved.

The first priority is usually to ensure that cash flow can continue after a server failure, and that all contractual obligations can be met during the period where the servers are being rebuilt. (An hour of lost transactions can mean thousands of dollars, and the loss of important accounts.)

  • At the very least, all of your most critical systems should have paper-based or off-line alternatives.
  • If this isn't possible, then server mirroring should be considered. This way, operations can be quickly switched over to another remote facility with minimal service disruption.
  • Another alternative would be to have a combination of remote and on-site backups. This allows for very fast recovery while still maintaining the safety of off-site data copies. If you choose to use this route for highly critical systems, make sure that the solution offers bare metal recovery and that it's optimized for extremely fast recovery speeds. (There are already a number commercially available solutions and services that can accomplish this quite well)

For cash-strapped IT departments, mirroring the entire datacenter might not be financially feasible. That's why an impact analysis should be performed beforehand. This will help in setting server recovery priorities, and assigning budget accordingly.

Another cost-saving approach would be to leverage the cloud.

By relying on the cloud for data protection, you can eliminate much of the up-front cost associated with implementing a high-availability plan. And because it's a pay-as-you-go service, cloud data protection allows you to stay current by modifying your disaster recovery plans as needed.

A data disaster is no time for panic and improvisation. By planning for the first seconds well in advance, you can greatly minimize downtime, data loss and expensive mistakes by ensuring a smooth and efficient recovery.

About The Author: Storagepipe has been offering cloud backup and other data protection services since 2001.

More Stories By Steven Rodin

Storagepipe Solutions are a leading provider of online data protection for business. Services include business continuity, email archiving, data archiving, and online backup.