Unsure about if outsourcing your DNS is worth it? Our latest tutorial on WebServer Compare defines some of the pros and cons of buying vs. building by using three examples of organizations with very different IT needs.
With all of the buzz and hype about the latest trend to outsource DNS service, there is much confusion on the topic. Is outsourcing really worth it? How expensive is it for an enterprise to run its own DNS server? Does a DNS Server create a lot of traffic? How intensive is DNS administration for the IT guys?
All of these questions raise pretty critical points, but they are also very difficult to answer without knowing the specifics about an enterprise.
In this tutorial, we will try to explain the pros and cons of outsourcing DNS management functions using three (general) examples of organizations with very different IT needs.
- Example No. 1 is a small office, home office (SOHO) or small business that has one server doing everything -- hosting its Web site, e-mail, files, and print capabilities. Its Windows 2000 server is running with 10 client access licenses. There is one IT professional on staff and one public IP address for the server.
- Example No. 2 is a midsize organization that has more than one server but not quite a server farm. It is running Unix and Windows NT/2000 on its servers and has claimed one or two public IP address for a few of its servers. The organization most likely has several full-time IT professionals on staff who look after the many workstations.
- Example No. 3 is a large enterprise with a dedicated staff of IT professionals running all different type of servers. It has a public IP address for every workstation and server as well as a large Server Farm occupying an entire area of building space.
The bottom line for any enterprise deciding whether to implement its own DNS service is that DNS itself doesn't create a lot of network or server load. DNS uses a simple database lookup and UDP packet system, some of the most basic things in the IT world.
Example 1: SOHO
A SOHO is so small that it is the most flexible when it comes to DNS outsourcing.
It also has the most to lose, as it is most likely just starting to grow. It probably has only one IT employee, and he or she might not even be full time. He might just be the guy in the office who built his own home computer, but doesn't have any IT certifications. Because he is only one person, he might get bogged down with the administration of the DNS, as well as the server and other computers on the network.
On the plus Side, Windows 2000 Server comes with a built-in DNS Server, so no extra software package need be purchased. This is a cost savings when compared to the monthly rate that would be paid with outsourcing. Also on the plus side, because the business most likely has only a few work stations and one public IP, the setup for Windows 2000 DNS should not be all that difficult. The designated IT employee can use the WINS capabilities of Win2K DNS to take care of workstations names, and she can set up only one host and a few aliases for the server for all public queries to use.
Example 2: Midsize Organization
A midsize organization, on the other hand, is an ideal candidate for an outsourced DNS. The IT staff probably already has its hands full with the servers and the user workstations; adding more server side administration will certainly not ease the burden.
A midsize organization is usually running both Windows 2000 and Unix, so it could either go with Windows 2000 DNS (easier to maintain) or Unix BIND (more customizable but requires someone with almost degree in Unix to maintain and work with it). Most companies that provide DNS services use a simple Web-based interface for updating DNS entries, which is much easier to use than the BIND version.
If a midsize enterprise does not wish to outsource its primary DNS services, it may instead decide to outsource its secondary services in case the primary DNS server goes down. For most midsize organizations, if users are unable to get to its servers via the web, it results in lost revenue in thousands of dollars per hour.
Example 3: Large Enterprise
Our third example is that of a large enterprise. It has back-up servers of every kind, or it is using clustering for its high-load systems, such as IMAP e-mail, file serving, and domain control. The enterprise already has a huge IT staff, and it's almost guaranteed that someone will be able to pick up DNS responsibilities with minimal training. More than likely, it will be the domain controller administrator who takes up this task, as the two services are very similar and can be intertwined in NT.
Outsourcing the DNS would be horribly inefficient in this case, as the enterprise has public IPs for each workstation and needs to dynamically update its DNS database as workstations go off and come online. Doing this with outsourced DNS services would be a waste of time and money. The only real way to have control over a large environment like this is for the enterprise to keep the DNS under its own control.
The following links provide more information about services related to DNS:Small and Free ServicesLarge Services With a Monthly Fee
This article was originally published on Friday Apr 6th 2001