You are on page 1of 10

40 Ways to Make

Your Data Center


More Efficient

an Server eBook
Contents…
40 Ways to Make Your
Data Center More Efficient

This content was adapted from ServerWatch.com. Contributor: Kenneth Hess.

2 10 Data Center Management Mistakes You Might Be Making


2
4 10 System Administrator Tasks Ripe for Automation

6 10 Free Server Tools Your Organization Needs

8 Uncover Your 10 Most Painful Performance Bottlenecks

4 6

8
40 Ways to Make Your
Data Center More Efficient

10 Data Center Management Mistakes


You Might Be Making
By Kenneth Hess

F
or those who think (falsely) that they have 3. Design Flaws
the perfect data center, read on for some Design flaws of a standing data center are difficult to
enlightenment. Those who work in the overcome, but a redesign is less expensive than a fresh
data centers of their dreams might beg build. A 20-year-old data center still looks good, but it
to differ with your fantasy. Though you may not achieve doesn’t perform up to today’s greener standards. You’ll
desired perfection affordably, you can come close by also have to retrofit your electrical apparatus to handle
changing the way you handle certain aspects of your data blade systems. You’ll probably need to toss that old
center management. Managing cooling system as well, since
a collection of computer systems contemporary servers run cooler
is no easy task. But, through and more efficiently than their
better management and proper predecessors did.
planning that task might involve
popping fewer pain pills. Here 4. Limited Expandability
are the 10 major data center “640K of RAM ought to be
mistakes to avoid. enough for anybody.” How many
times have you heard that quote
1. Inadequate Virtualization that’s attributed to Bill Gates,
If you operate a data center circa 1981? Whether he said it
and haven’t caught on that is of little importance now. The
virtualization saves money, lesson to learn is that when you
you’re way behind the curve. build anything, pretend you’re
Virtualization saves valuable rack converting a Celsius temperature
space. It saves additional money to Fahrenheit: Double the amount
on cooling, power and service you think you need and add 32.
contracts for those non-existent Using the Celsius-to-Fahrenheit
systems. equation will allow for some expandability in your data
center. Two thousand square feet of floor space isn’t
2. Untapped Cloud Computing enough? Try 4,032 feet instead. Poor planning is no
Similar to virtualization, cloud computing requires that reason to run out of floor space or any other capacity.
you obtain a clue about its capability for your company or
your customers. Amazon.com offers flexible and scalable 5. Relaxed Security
plans that fit into an on-demand capacity scenario. Using Enter any data center and you’ll see card readers,
Canonical’s Ubuntu Linux Server Edition, for example, retina scanners, circle locks, weight scales or other high
you can create your own private cloud or leverage technology security systems in place. But, next to those
Amazon.com’s Elastic Compute Cloud (EC2) dynamically. extreme security measures, you’ll see a key entry access

2 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.
40 Ways to Make Your
Data Center More Efficient

door for security bypass. Physical security requires no Fahrenheit, you’re wasting money. Servers need air flow
bypass. If there’s a bypass in place, consider your security more than they need arctic temperatures. Take a stroll
compromised. through your data center. If it’s comfortable for you, it’s
comfortable for your servers. There’s no need to freeze
6. Haphazard Server Management your data center employees or make them sweat.
To manage your server systems, do you need physical
access or can you manage them remotely? Every 9. Underpowered Facility
contemporary server system comes with a maintenance How many times have you heard that a particular
connection with which to manage that system remotely. data center has floor space but no more power? You
Use it. Enable it. For each person who enters a data hear it more than you should, if you hear it at all. An
center, you can expect some amount of system failure. underpowered facility is a victim of poor planning. (See
Incorrectly labeled systems, incorrect locations, a misread No. 4 above.) Virtualization can help give you back
system name — the list goes on. Do yourself a favor: some power. Server consolidation can also assist. But
Enable those remote access consoles when you provision those are short-term fixes for the greater problem of an
your physical systems. underpowered facility.

7. Ill-fated Consolidation Efforts 10. Rack Overcrowding


One order of data center management business is to If you’ve ever attempted to work in a fully populated rack,
minimize the number of systems on the floor or in the you probably wished you had miniature hands or extra
racks. Server consolidation is the method by which this long fingers. It might seem inefficient to leave a bit of
effort is carried out. Consider a consolidation ratio of space between systems, but those who have the job of
2-to-1 or 3-to-1 unacceptable. Physical systems that plugging and unplugging components for those systems
operate in the 5 percent to 20 percent utilized range can will thank you. Poor planning leads to rack overcrowding,
easily consolidate onto a system with five, six or more of and it’s unnecessary. Virtualization, consolidation and
its peers. Underutilized systems waste rack space, power a more efficient arrangement will ease the problem.
and money in the form of service contracts. Experiencing an outage because of accidentally
8. Overcooled/Undercooled Space unplugging a server might convince you to leave a bit of
What temperature is your data center? You should find space between systems.
out. If your data center operates below 70 degrees

3 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.
40 Ways to Make Your
Data Center More Efficient

10 System Administrator Tasks


Ripe for Automation
By Kenneth Hess

A
system administrator who does everything groups, but when it comes down to real management,
manually wastes not only her time but yours you probably haven’t found the right tool. The one you
as well. Tasks that a sys admin performs want might have to be the one you create yourself via
repeatedly should be automated. Automation scripts.
through scripting, specialized software and system
scheduling frees her time, saves you money and prevents In UNIX, it’s simple to create scripts to prompt you for the
mistakes due to human error. These 10 sys admin tasks accounts that you wish to remove, have the system copy
are prime targets for automation the user’s files to a new location,
and will help streamline your daily change the permissions, search all
operations. systems for any files owned by that
user, change permissions on those
1. Patching files, or move them and complete
The only time manual patching is the process by removing the
called for is when that stubborn user account from the directory
minority of systems will not take service. Check out some of the
patches by automated means. add-on modules for your user
Linux and Windows include tools management tool of choice.
to perform automated updates, Microsoft, for example, offers its
but if you’d like more control Active Directory Resource Kit Book
of which patches your systems and CD that includes utilities for
receive and when they receive automation scripting.
them, investigate HP’s Data Center
Automation Center (HPDCAC) 3. Security Sweeps
software (formerly Opsware). Much more than just You should perform regular, automated security sweeps
an automated patching application, HPDCAC moves on your entire network to expose and fix any wire-borne
managing a complex infrastructure into a single, simple vulnerabilities. The frequency and intensity of the scans
interface. If you’re looking for a “patch only” solution, depends on the complexity of your network. Through
check out Ecora’s Patch Manager for agentless patch scripting magic, you can set up scheduled scans, send
management. the output to a database, extract a post-scan report from
the database, and email it to yourself or create an HTML
2. User and Group Maintenance version of the report suitable for online viewing. One
You’ve probably used Active Directory, LDAP, NIS+ or such tool, available for every modern operating system, is
other user and group account management software, but Nmap. Nmap is a free network security scanner designed
have you ever used one that really made you happy? The to rapidly scan large networks and report vulnerabilities.
reason you haven’t is that there’s not a lot of automation
built into them. Sure, you can create a user account, 4. Disk Usage Scans
remove a user account, and create groups and manage There is a constant turf war raging between users and sys

4 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.
40 Ways to Make Your
Data Center More Efficient

admins, and it is one that the sys admin must ultimately campaigns and other time-specific events. Moving the
win. To that end, the sys admin has some tools to employ: code from one environment to another manually is
disk space quotas, disk partitions and disk space scans. cumbersome, error-prone and requires coordination
Scans are regular audits of disk space usage by user. between developers and sys admins. Enable your
Offenders usually receive a warning or two before developers to promote code from one environment to
personal contact from a sys admin. Typical remedies for another using an automated code deployment system.
disk space gluttons are temporary account suspension, Some sys admins use RSYNC for automated code
removal of files, moving the files to a new location or an deployment and it’s safe to use if coupled with SSH keys
extension of the user’s space quota. These automated to secure the transfers between hosts.
scans, when performed regularly (about once per week),
prevent harsh actions by the sys admin and keep users 8. High-Level Administration
apprised of their disk use. You can perform those housekeeping duties, service
restarts and maintenance notices through automation.
5. Performance Monitoring Set up your scripts to fire during low-use hours for
Taking an occasional performance snapshot is a good clearing temporary file dumps, restarting your favorite
method for a single point-in-time glance at system services and sending out any maintenance or downtime
performance. That singular peek is only a pixel in the notices via email. You’ll find that automating these tasks
entire performance picture. You need something with takes some of the pressure off of you to remember which
more depth and breadth that will provide you with day it is and which list of things you need to do. There’s
performance trends and predictive peaks and valleys. no reason to keep a calendar of these; let the system
Setting up such a system is easy with Orca. Orca compiles handle them.
performance data from disparate sources (UNIX,
Windows, Linux) and creates easy-to-read performance 9. Reboots
graphs. Gathering of data, calculations, graph generation Yes, you can automate system restarts. Sitting around
and display are all part of the automated system. waiting for systems to bounce back to life is a waste of
time. Automate the process during low-use hours. Don’t
6. File Transfers worry, your automated monitoring system will notify you if
Using command-line scripting power (Windows, UNIX the system doesn’t come back online within a reasonable
and Linux), you can perform automated file transfers amount of time.
between hosts. There’s no need to do them interactively.
If you’re clever in your timing, you can set up elaborate 10. Malware Scans
automated schemes that not only transfer your files but You can scan for spyware, malware, viruses and other
also unzip, change permissions, move, copy and insert nasties using automated processes. Using scripts, you can
information into a database. Use the secure versions map or mount drives, scan your filesystems, disconnect
of your file transfer utilities (e.g., SSH, SFTP, SCP) to when finished with the scan, scrape the scan log for
ensure that anyone snooping doesn’t grab an important positive hits, and send the results to a database or in an
password from your network stream. email. You don’t need to manually perform these scans
when your system is perfectly happy and suited to do so
7. Code Promotion on its own.
How you promote code from test to staging and into
production can have a profound effect on marketing

5 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.
40 Ways to Make Your
Data Center More Efficient

10 Free Server Tools Your Organization Needs


By Kenneth Hess

T
his list of 10 free, essential tools is an 3. Nagios
amalgam of tools for all sizes of companies Nagios is an enterprise infrastructure monitoring suite.
and networks. The range of tools covered It’s free, mature and commercially supported. It has
here are generally cross-platform (i.e., they grown from a niche software project to a major force in
run on multiple OSes) but all are extremely useful to the contemporary network management. It’s used by such
system administrator, network administrator and first- high-profile companies as Citrix, ADP, Domino’s Pizza,
level support personnel. While all of these tools are free Wells Fargo, Ericsson and the U.S. Army.
to download and use in your network without payment
of any kind to their developers or maintainers, not all are 4. Wireshark
open source. The 10 essential If you run a network of any
tools listed here, in no particular size or topology, Wireshark is
order, are from various sources a must-have application. It is
and represent the very best in a network packet capture and
tools currently used in large and analysis program that assists you
small enterprises alike. with your ongoing quest for a
trouble-free network. Wireshark
1. PSTools won’t prevent network problems,
PSTools is a suite of useful but it does allow you to analyze
command-line Windows tools those problems in real time and
that IT professionals consider possibly avoid failure.
essential to survival in a
Windows-infested network. It 5. Apache
provides automation tools that The Apache project isn’t just a
have no rival. There is no greater web server. The project, officially
free toolset for Windows available known as the Apache Software
anywhere. Microsoft provides this Foundation (ASF), consists of
suite free of charge. If it’s not part almost 100 different projects
of your Windows diagnostic and automation arsenal, stop under the Apache umbrella. Yes, the famous and wildly
reading and download it now. Be sure to come back and popular HTTP server, Apache, is the project’s namesake
finish the list. (You can multitask, can’t you?) and mainstay, but it isn’t the only nymph in the forest.

2. SharEnum 6. IP Plan
ShareEnum is an obscure but very useful tool. ShareEnum IP Plan is a little-known project that has potential in any
shows you all file shares on your network. Even better, it size environment. It’s not a DNS service, but it is a Web-
shows you their associated security information. This very based, IP tracking application. The reasoning behind a
small (94K) tool might become one of the most valuable tool like IP Plan is that DNS tracks systems that are in use.
and useful security tools that you possess. It is another But to whom do you go when an IP address conflict, and
free tool from Microsoft.

6 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.
40 Ways to Make Your
Data Center More Efficient

how do you know which IP addresses are free to use? You 9. OpenOffice.org
won’t — unless you have a tool like IP Plan. It’s easy to OpenOffice.org (OO.o) is the free equivalent of
use and free. What more could you want? Microsoft’s popular office suite. OO.o sports a word
processor, spreadsheet, presentation program, database
7. Eclipse and more. It is compatible with Microsoft Office and can
Eclipse is an Integrated Development Environment use or export in almost every imaginable file format.
(IDE), which you can use to create applications with OpenOffice.org is not only easy on the wallet (free), but
almost any computer programming language. Eclipse it’s also the darling of IBM, which has created its own
has wide language support, but it is historically viewed derivative: Lotus Smartsuite.
as a Java development tool. You can develop Windows
applications in this very complete IDE as well as 10. Webmin
applications for every current operating system. Webmin, for the uninitiated, is the ultimate lazy system
administrator tool. It’s a Web-based interface to
8. KVM your UNIX or Linux system that covers almost every
Kernel Virtual Machine (KVM), now owned and supported configurable aspect of the system and any add-on
by Red Hat, is a free, full virtualization solution. Full program you can ponder. You can’t rely on it for 100
virtualization means hardware abstraction enables you percent of your system administration tasks, but you can
to use almost any OS in a virtual machine. Each virtual probably use it for 99 percent of them.
machine has its own display, network, disk and BIOS, and
it functions like a physical system. You install an OS into
a virtual machine just as you would to a physical system.
Yes, even Windows.

7 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.
40 Ways to Make Your
Data Center More Efficient

Uncover Your 10 Most Painful Performance


Bottlenecks
By Kenneth Hess

W
hen you hear the words, “performance other system flaws that manifest themselves as memory
bottleneck,” the typical hot spots that come issues. The key to solving memory performance problems
to mind are CPU, Memory, Disk and Network. is to find the root cause of the symptom before adding
Those are good places to start looking for more RAM.
bottlenecks but they aren’t the only places performance
problems can hide. This list targets six other potential 3. Storage
leads for your investigation into the elusive performance Disk speed, RAID type, storage type and controller
breakdown. Sometimes just knowing where to look technology all combine to produce what’s known as
might prevent your own personal disk I/O. Disk I/O is a common
breakdown. source of performance angst
Note that listed items are in no for system administrators
particular order. and users alike. There are
practical and physical limits to
1. CPU performance even when using
The CPU is the brain of the the best contemporary disk
computer where calculations and technology. Use best practices
instruction operations occur. when combining and separating
CPUs can handle millions of workloads on disks. As attractive
calculations and instructions, as leveraged storage is, local
but performance suffers when disks are still faster than the
the numbers of these operations fastest SAN.
exceeds capacity. CPUs that
sustain greater than 75-percent- 4. Network
busy numbers will slow the entire The network is a commonly
system. CPUs need some room blamed source of performance
for activity “bursts” where loads can reach 100 percent bottlenecks, but it is rarely
for short periods of time. CPU load is a common source found to be so. Unless there is a network component
of performance bottlenecks. hardware failure, such as a damaged switch port, bad
cable, jabbering network card or router configuration
2. Memory problem, you should look elsewhere for your “network”
continued
The rule of thumb on memory is “add more.” When performance bottleneck. A perceived slowness on the
performance problems point to memory, the general network usually points to one of the list’s other nine
consensus to solve the problem, is to add more. This entries.
practice is effective only in the short term, however.
Performance bottlenecks that point to memory are often 5. Applications
the result of poorly designed software (memory leaks) or Although no application developer wants to hear it,

8 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.
40 Ways to Make Your
Data Center More Efficient

poorly coded applications masquerade themselves as 8. Failing or Outdated Hardware


hardware problems. The fickle finger of guilt points to The older the hardware, the more likely it is to fail.
applications when an otherwise quiescent system suffers Some hardware components fail with a single final
greatly when the application is on and shows no signs breath, while others linger on with random complaints
of difficulty when the application is off. It’s an ongoing and untraceable glitches. Hardware that causes system
battle between system administrators and developers reboots, disappearance of data or performance
when performance issue occur. Each wants to allege the bottlenecks frustrate system administrators because of
other’s guilt. A word to the wise after many hundreds of its unpredictable nature. The best way to prevent such
hours of chasing hardware performance bottlenecks: It’s tragedies is to keep hardware fresh, use redundant
the application. hardware and monitor your systems carefully.

6. Malware 9. Filesystem
Viruses, Trojan horses and spyware account for a large Did you know that your filesystem choice can have
percentage of perceived performance bottlenecks. Users a profound impact on performance? It can. Some
notoriously complain about the network, the application filesystems, JFS for example, uses very little CPU. XFS
or their computer when nasties raise their ugly heads. has very high scalability and high performance. NTFS is a
Those performance killers can reside on one or more recoverable file system with high performance. The new
server systems, the user’s workstation, or a combination EXT4 filesystem supports very large files efficiently. Each
of the two. Malware infections are so common that you filesystem has a purpose, and using the incorrect one for
must employ multiple defenses against them. Antivirus, an application can have disastrous results. Consider your
antispyware, local firewalls, network firewalls and a filesystem choices wisely and select the best one for the
regular patching regimen will help protect systems and job. There is no one size fits all filesystem.
prevent resultant bottlenecks.
10. Technology
7. Workload The technology you select for your infrastructure plays
Smart workload management can help prevent an important role in performance. For example, if
performance problems associated with poorly balanced you dedicate your services to a virtual infrastructure
workloads or ill-conceived load balancing schemes. technology, you might have performance problems
Adding another system to a suffering cluster relieves the not experienced on equivalent physical systems.
pressure, but this is easier to do in a virtual environment Alternatively, there are some workloads that thrive
than in a physical one. The best advice here is to measure on virtual technology. LAMP (Linux, Apache, MySQL,
capacity and performance of all systems and heed the PHP) workloads, for example, perform at and greater
numbers reported to you. Move workloads, add systems than native speeds on KVM. However, container-type
and keep a watchful eye on performance. virtualization (OpenVZ, Parallels, Solaris Zones) boasts
native performance ratings for any workload.

9 Back to Contents 40 Ways to Make Your Data Center More Efficient an Internet.com Server eBook. © 2011, Internet.com, a division of QuinStreet, Inc.

You might also like