Archive for the ‘Storage’ Category

Recently, I’ve had the experience of purchasing and attempting to get a NAS device working on my network. The Seagate Black Armor was recommended. My experience with the device unfortunately has not been rewarding.

The Seagate Black Armor has been noted to support up to 4x3TB drives. That is a total of 12TB.

Upon opening the box, setting up the device, and inserting the 4x 3TB Seagate harddrives, the box ran smoothly and the lights lit up in green with “Black Armor” turning up on the LCD screen. The tiny instruction booklet with brief instructions said that the drives should take 8-9 hours to be prepared before the words “Black Armor” turned up on the LCD. It turned up in less than 10 mins.

So maybe it was I had been playing around with the harddisks. I had attempted to get it to work on a Redhat workstation and used fdisk/parted to manipulate the drives partitions.

After querying the LCD screen for the IP address the NAS device had adopted, I open a web browser to view/configure the NAS device. Access to the NAS device via a browser was positive.

A simple login and TADA, the drives were viewable. However the NAS device has not properly accessed the 4x3TB devices. They were just sitting there and creating access volumes was a challenge

Seagates’ advice was to reset the Black Armor device. This is done by clicking on the pinhole at the back of the device. The NAS device lights would turn amber, and the device would finally reboot.

One of the biggest problems of delivering value in a business intelligence project is providing insight around a dataset. Delivering insight about any particular dataset is not about successfully processing the data in question and analysing it. In today business intelligence (BI) world, the expectations are alot higher. Valuable insight is derived from co-relating a particular dataset with sometimes a very different abstract perspective/dataset.

An Example

You have a dataset on radiation levels. (thanks to fallout from nuclear powerstations). A very quick and common question that demands immediate answers would be “What is the impact of increased radiation?”. That is a very broad question, and even with skillful narrowing of the scope of the question, this question still needs to be answered. Even the basic remaining key perspectives on the question may be:

Effect on population? Effect within a radius of 100km? Effect on transportation within 100km? Effect on travel? Effect on tourism? Effect on agriculture?

All these questions will require the custodians of co-related datasets to make their data available. The negotiations to acquire the data would probably take time. Followed by the data modeling, loading and analysis. The final outcomes would still be achieved, but under the strain of time and effort.

We can reduce some of this time by having open data, and configured data. Consider plug and play data. Consider being able to draw data from established datasets with minimal processing, and be able to derive results quickly. This is where Glitchdata would advocate data by convention.