When Atlantic.Net was asked to help answer some questions for an article to be published in the computer world, we turned to Tony S., our Sr. Windows Administrators. With his 15+ years of full spectrum IT experience, Tony is a Microsoft specialist and also holds certifications from Cisco, Citrix, and VMware.  He has experience in development of advanced technology solutions and designing/building/managing IT operations for both internal and customer-facing systems.

Tony explains how the Internet is “broken” and what we all can do to help fix it!

How is the Internet broken today?

I don’t believe the Internet is broken, per se. It’s just experiencing some impressive growing pains. When the Internet was originally designed, the creators had no idea what it would eventually become and did not design it to do even one-tenth of what it is capable of doing today. Technology professionals are no longer focusing on repairing the existing shortcomings and are instead looking to build for the long range future. IPv6 is a good example of this.

What is happening technically when news about a celeb like Michael Jackson hits and causes a lot of server failures? It is more than just “a crashed server”. Why can’t servers keep up from a technical standpoint?

I’d be tempted to call this poor planning but really it’s the effect of a balancing act that technology professionals have to constantly perform. Approaching your CEO with a request for 50% more infrastructure simply because at some time in the future demands may go up by 200% for a few days doesn’t make sense from a financial perspective. Businesses design their networks to meet the needs of what could be called ‘average’ utilization with a bit added in for growth. Anything beyond that is normally considered overkill.
A new trend like Managed Server Hosting allows companies to scale their capacity with a simple phone call to their provider vs. the old method of waiting weeks for servers to arrive from a hardware vendor.

What are four or five ways to fix the Net so that a site is always up no matter how much traffic? This can be related to bandwidth, HTML, virtualization, etc.

Continue to develop bigger and faster data transfer technologies. Continue to address inefficiencies in current technology. Design infrastructure with focus on load balancing and fault tolerance at the forefront.

Can you describe a technical scenario where tens of millions of visitors use a site like Twitter, a news magazine, or Facebook and it stays live? What are the steps to ensure that the site stays up ?

The most important step is to host servers in a secure and robust Data Center; this allows ample capacity available on demand. We have a client who appeared on the Oprah Show and wanted to make sure their site would stay up for the very same concerns. We were able to help them out by allowing them to burst literally up to an infinite amount of bandwidth by utilizing multiple carriers in our carrier neutral facility. Data Centers generally have multiple layers of redundancy and fault tolerance built into their infrastructures for mission critical sites. Security is another area where expert engineers can implement security measures so that the site is not vulnerable to DDS attacks and others. Electric power is another area where a lot of newer technologies require dedicated power resources that are only available in a dedicated environment with multiple connections to utility power backed by industrial grade UPS’ and generators. I believe the scalability of new technologies such as cloud computing (although still in its infancy) will contribute to the success of several large traffic sites (social networking, etc).

What are some roadblocks to fixing the Net?

It depends. For instance, we as a Data Center operator have to overcome many roadblocks on a daily basis to ensure service levels are met. Sometimes we have to think outside the box and come up with solutions that not only solve the problems but make long term sense too. Also the Internet industry has gone through massive consolidation over the last 15 years, so you have less parties innovating where infrastructure is concerned. On a larger scale though, the major roadblocks are cost and time and they always will be…however, smart operators, like Atlantic.Net, will always find a sensible way of making it work.

In terms of innovative stop gap measures, what can work and what is more wishful thinking?

Cloud computing can work. Virtualization works and works well. IPv6 will work (whether you’re ready for it or not). Expecting wired network performance without the wires still isn’t happening. I’ve not yet see the ‘magic bullet’ that will address this issue. The Net isn’t broken, just our expectations.

Tony’s  suggestions to fix the “broken” Internet:

1. More bandwidth equals more throughputs. Bring on the 100Gbps switching! Keep in mind if you must have more bandwidth, you must pay more to get it.
2. CDN’s have made a huge impact on content delivery. They take the load of off other systems and allow better overall performance and scalability.
3. Caching and hardened DNS won’t fix the Net, but helps from a security standpoint.
4. Redundancy is always good.
5. Never accept failure as option, only as an obstacle to be overcome.

Tags: Cloud Computing, Colocation, Data Backup, Data Center, Developers, Managed Dedicated Servers, Managed Hosting, Managed Servers, Security, Server Hosting, Support, Tony Segner

This entry was posted on Thursday, November 5th, 2009 at 10:18 am by jeisenberg and is filed under Data Center Services. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.