@SmallBizCIO:
A blog aimed at the SMB, written by a long time IT enthusiast turned IT professional.

"Jack of all trades, master of none,
Certainly better than a master of one"

These are my personal opinions.

  1.  

    High Availability in the SMB

    When I use the words High Availability (HA) when talking to my clients, it usually results in funny looks and them telling me that their environment is not a 24 hour shop. Asking further, HA is often equated with large expenses in purchase as well as maintenance.

    On a separate note and just to be clear: Having backups available is NOT availability (high, low or anything in between). Backups are for catastrophic loss of data and are not meant to provide ongoing service when failure occurs.

    Let’s get a few terms out of the way. At the core, there are 3 parts to keeping a service alive throughout a failure: redundancy, fail-over and high availability. 

    • Redundancy is a ‘physical’ element, providing multiple units of the element in case one fails (servers, services, connections, etc.) 
    • Fail-over is a process which provides for the transfer of functionality to the redundant element. This requires an ongoing synchronization of data between the elements. Fail-over can be done manually or automatically.
    • High Availability is a measurement, most often associated with the time in which a fail-over takes place. In order to maintain high availability, the fail-over is usually done automatically, based on certain indicators.

    All three go hand in hand and more advanced technologies such as clustering all ultimately operate on this principle.

    The big question often becomes; what part of the infrastructure needs to be highly available? And, in many cases, I agree that most of the infrastructure in the standard SOHO environment is not required to have HA.

    However - Think about your firewalls. What happens if you loose your Internet connection? Are you relying on email? Are you operating in the Cloud? How much does it cost YOU when your provider drops the ball?

    If your budget is low and you still want an Internet connection that is up, mitigating as much risk of downtime as possible, consider the following option:

    1. Get 2 internet connections through different technologies. DSL or a T1 as the first technology and a cable or WiFi connection on the other.
    2. Get a router/firewall that can handle the 2 connections and that can automatically switch traffic if one connection fails. Examples of solutions that can do this are pfSense, Fortigate 60 Series and the Cisco ASA 5500 Series. These devices range from $250 - $750 with the proper licensing and configuration. Highlight goes to the pfSense because it is Open Source and can be ran on any type of hardware - even on an older computer or server to test it out.
    3. For the price of the router, get a second one and configure it in automatic failover mode. 

    This setup usually costs less than $1000 up front and adds $100-$150 for the secondary Internet connection (YMMV). Compare this cost with the total cost of being down for a few hours, a few times a year (which seems little here in South Florida) and it will earn itself back within the first year.

    Next up - HA on your database servers.

  2. blog comments powered by Disqus