Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

  1. Home
  2. Networking
  3. Building-reliability-it
  4. Page
  5. Building Reliability Into IT: Page 2 of 2
Networking

Building Reliability Into IT: Page 2 of 2

There's a lot to like about Twitter; for me, the interesting discussions that crop up are a main draw. This was especially true recently, regarding a conversation about the foundations and premises of cloud computing. The discussion (sorry, I have yet to find a way to retrieve an entire Twitter thread) started with Sam Johnston saying "Legacy: unreliable software on reliable hardware. Cloud: reliable software on unreliable hardware." That's the kind of loaded statement Johnston likes to throw o
Mike Fratto
September 27, 2010

Many custom built applications are not good candidates for parallel reliability because the original specifications were modest and relied on expensive hardware for reliability. Informationweek Analytics contributor Mike Davis pointed out in a conversation with me that retrofitting existing applications may not be as hard as you think. Making the software smarter may not require a complete rebuild of the application. Any modern software application should be written in a modular enough fashion that separating the tiers, for example, and adapting a service broker model, can make the application more amenable to parallel reliability.

Regardless of the method used to modularize applications, there will be an impact on the network design and operation. More transactions will require lower latency networks to improve application performance. More network equipment like ADC between application layers is going to have to be factored into overall application designs and data center operations. The network layer is going to have to be more aware of application location. The separation between application architecture and network architecture is going to get very blurry.

It's not going to be enough for the network team to design the network in a vacuum. Similarly, the application development team is going to have to learn more about the capabilities built into the network. Both teams are going to have to work together to ensure your organization runs reliable applications. You can look at this additional team work as a chore to be completed or as an opportunity to build better systems. As IT professionals, I'd think the latter is the preferred choice. 

A private cloud can play a key role in your disaster recovery strategy.
We dig into the storage, LAN, and WAN requirements to build a cloud for
DR. That and more--including articles on automated data centers and SaaS
Web security--in the new all-digital issue of Network Computing. (Registration required.)

Tags:

Commentary
Networking