Infrastructure Best Practices
September 15, 2009
We’ve been doing a good bit of consulting lately with accounting firms of various sizes. Initially, we would start these conversations more or less assuming that the core pieces of each firm’s network were done according to accounting industry best practices. However, we learned quickly that this is not always the case. In light of this finding, we would like to share with you a few things that jump out as infrastructure best practices for a network:
- RAID – make sure that your servers have a RAID configuration on both the OS and Data partitions. For example, if you have a File & Print server, you would want to have your OS partition in a RAID configuration, as well as your Data partition in a RAID configuration.
- Redundant power supply – this one is fairly self explanatory. Redundant power supplies are a fairly standard option on new servers these days. However, if your server is a couple of years old and does not have redundant power, it would be something worth looking into.
- Dual gigabit NICs – gigabit NICs are usually a given on a server, however dual NICs are not always as common. Xcentric always implements dual NICs on servers. You can run these in a failover mode or as load-balancing.
- Gigabit switches – you want to have at least all of your core network servers connected to a gigabit switch to reduce any bottlenecks.
- Server warranty – it is always in your best interest to have enterprise-grade production servers that are covered with some sort of next business day service & support contract. You may have lasted the previous 3 years without one, but that is a big risk to keep taking.
- Server room- Computer equipment may look tough on the exterior; however, it is very sensitive to the environment that they operate in. The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) recommends a temperature range of 20–25 °C (68–75 °F) and a humidity range of 40–55% with a maximum dew point of 17°C as optimal for data center conditions, according to their “Thermal Guidelines for Data Processing Environments”.
- Power source- A single server can draw up to 300 Watts of power, which means special planning must be made to supply uninterrupted (backed up) power. Datacenters and server rooms should not share power circuits with other high-power consumers such as copiers, refrigerators or air conditioners. Some battery backup units also require special outlets to be installed from dedicated power circuits.
- Server hardware- Only Tier One hardware should be used for the production environment. Your entire firm has its lifeline routed throughout this hardware and our best practice is to use either Dell, HP, IBM, or severs of this caliber.
- Server patches and updates- Operating systems are full of security flaws and bugs that need to be patched on a regular basis. If these critical updates are not installed, the system can be vulnerable to attacks and data can be compromised.
- Remote access card (optional) – Dell has a card called a “DRAC” (Dell Remote Access Card), and HP has something similar. These are virtually like another NIC, however, they allow you to gain access to the server, even if it is turned off or inaccessible via the network. This is a very valuable tool.
If you have any questions regarding infrastructure best practices, feel free to contact us at 678.297.0066 or support@xcentric.com.
2 Comments
leave one →
Michael
Great post. Based on our discussions, I think it is often the case that firms 10-20 do NOT have most if any of the best practice solutions you discuss and yet all of them dramatically impact performance and reliability
Great job
Frank
this is just the bare minuim, it’s scary to think they dont even meet these requirements.
good article
-JOSH