New Server Designs Lower the Cost of Enterprise ComputingBy Michael Vizard | Print
Re-Imagining Linux Platforms to Meet the Needs of Cloud Service Providers
Server designs are going through a radical transformation that should have some far-reaching impact throughout the channel. The best examples of this transformation come in the form of new server designs that put multiple functions and services into a single box using blade server technology.
Server designs are going through a radical transformation that should have some far-reaching impact throughout the channel.
Two of the best examples of this transformation come in the form of new server designs from Hewlett-Packard and Critical Links.
HP this week has rolled out a new Adaptive Infrastructure in a Box for Midsize Business offering that combines what were once several separate servers into a single box running Windows Server 2008.
Based on a blade server architecture, the core idea behind these servers has been a long time in coming from when vendors first started rolling out blade servers. What HP has done is dedicated specific processors inside the blade server to handling specific functions in order to eliminate the need for specialized server appliances.
Similarly, Critical Links has leveraged open source technologies to create an all-in-one EdgeBox network server that combines telephony, firewall, router, wireless access point, file and print server, web and electronic mail server, unified threat management and traffic prioritization technologies in a single server.
Both approaches represent a potential for substantial savings for customers not only in the cost of acquisition for server technology, but in the number of people it takes to manage various servers.
Of course, some solution providers might bemoan the fact that what we are really talking about here is selling fewer servers, which might have an adverse impact on profit margins should a solution provider wind up selling fewer actual servers.
But for any solution provider that has a business focused on managed services, the advent of all-in-one servers could represent a substantial savings on customer service. Instead of having to master multiple systems management systems, these servers can be remotely managed using a common console. That’s should result in less cost in training and the actual number of people needed to manage these systems.
The vendors, of course, will also argue that these designs also represent an opportunity for having an upgrade conversation with a customer in a way that ultimately saves the customer money on IT staffing. In effect, this means that these systems can almost pay for themselves.
The HP offering has a starting price of about $18,000 for a system that can support up to 300 users. The Critical Links server is roughly $3,000 for an implementation that supports up to 40 users, which it compares to a similar implementation based on Microsoft technologies that would cost $9,673.
Odds are strong that these servers only represent the first in a long line of new server designs. As multicore processors move mainstream, solution providers should expect to see a whole new range of servers that leverage different processors to perform specific functions.
The ultimate end result of all this work should be the end of server sprawl over the IT infrastructure. And that’s a good thing because we all know that the real cost of enterprise computing is not the products themselves, but rather all the people we need today to manage it.