Enable the network

The Domain Name System (DNS) is used by practically every device on the network as a way of locating other systems. In a corporate environment, it is typically end user devices such as PCs, laptops, and tablets that are sending DNS queries in order to resolve the names and IP addresses of application servers, but behind the scenes, application services located in data centres will also be performing massive numbers of DNS “reverse” lookups to convert the IP address of each incoming connection into a name, so it can be seen that DNS is used by both ends of the client-server communication. A failure at either end can lead to application delays, connectivity problems and authentication failures. In fact the number of failures that can be caused by a DNS issue is varied and can be very difficult to troubleshoot without intimate knowledge of the application involved and the way it uses DNS.

With IPv6, IP addresses are so complicated that it is impossible to remember them, meaning that the reliance on DNS will only every increase.

As corporates embrace BYOD and IPT, there are more and more devices reliant on DNS so it is essential that the DNS servers are sized correctly with sufficient capacity and that they are configured in a resilient configuration, such as in High-Availability pairs or implementing Anycast.

The Dynamic Host Configuration Protocol (DHCP) is used by devices to obtain a valid IP address from your network. Every time a smartphone or iPad comes onto the network, it will obtain an IP address from a DHCP server. If you have embraced BYOD, DHCP is critical to the success of this as not only does DHCP provide the correct IP address, it also provides other important parameters such as the IP addresses of the DNS servers that should be used.

With many companies deploying IPT, the VoIP handsets are also totally reliant on DHCP to obtain their IP address and configuration information automatically when they boot up. When sizing a DHCP server, it is important to remember the VoIP handsets when looking at capacity and performance statistics. If you operate a call centre that has 10,000 handsets and suffer a power glitch causing all the phones to reboot, can the DHCP server handle that many phones requesting IP addresses all at the same time?

DHCP can be configured in a failover configuration whereby more than one DHCP server can share the same address pool. This enables the load to be spread across multiple servers whilst also providing resilience should one server be unreachable.

The Network Time Protocol (NTP) is used to distribute accurate time signals to systems on the network. If you do not currently have NTP deployed, then the clocks on each of your systems will be wildly inaccurate and unsynchronised.

Time synchronisation is not only important for auditing and compliance reasons, but many networking features require an accurate clock, including DHCP failover and DNS security.

There are many publicly available NTP servers available on the Internet that are available to query for free, but some organisations require incredibly accurate clocks or do not want to rely on the Internet for their time feed. In these cases it is recommended to implement a dedicated NTP appliance that synchronises time from another source, such as the GPS satellite network or long-wave radio service operated by the National Physical Laboratory.

The Trivial File Transfer Protocol (TFTP) is an often overlooked file transfer protocol but is often used as the delivery mechanism for transferring binary/boot images to VoIP handsets when they boot. Typically a VoIP handset will boot up and request an IP address via DHCP. Part of the DHCP packet will contain the IP address of a boot server and file name of the boot image; the handset will then communicate with that server once it has an IP address in order to retrieve its boot image. The boot image contains the firmware that the handset needs in order to operate, and so TFTP forms another critical part of the infrastructure.