Gabriel consulting whitepaper on enterprise IT for SMB

Enterprise IT Management for the SMB?
Keeping a SMB running smoothly is like walking a tightrope. SMBs depend on IT
just as much as larger organizations, but don’t have the skilled personnel and
management tools to ensure they have a highly available infrastructure that
delivers predictable performance. System and application management in the
SMB is mostly firefighting - fixing critical problems when they arise – and often
consumes the IT staff entirely. IBM’s Tivoli unit thinks it's found a way to give
SMBs what they need, with a learning curve and price tag to match. The IBM
Smart Business appliances can help restore order and prepare a business for
integrating state-of-the-art, Web 2.0 technologies without dramatically
expanding the existing environment nor sending employees for weeks of
training.

As a technology analysis firm, we’d argue that technology is one of – if not the – most important
tools used by today’s successful businesses. Computing technology is entrenched. We use it not
only to count and track things in the enterprise (traditional back-office data processing), but also to
leverage the value of highly skilled, front-office employees. Personal productivity gains driven by
IT, help people do far more today than they did even as little as a decade ago. Information
Technology also serves as a main means of business communication: the corporation conducts
internal conversations via email and instant messaging, and can quickly and easily reach out to
market, sell, and service their products.

All of these capabilities are typically taken for granted in most Fortune 500 organizations, but
making them work and keeping them working is far from a trivial task. Large enterprises have tens
of thousands of servers (many virtualized) running huge numbers of applications, each of which is
important in some way to accomplish the goals of the organization. They have large IT staffs chock-
full of specialists who focus on systems, networks, and storage, and are dedicated to specific
applications.

Driven by the Worldwide Web and numerous Web 2.0 technologies like wikis, blogs, and social
networking sites, IT usage in terms of server and application counts continues to explode, and the
need to have every application and process constantly available is a given. It’s not ‘ok’ to have the
corporate website down for a day or two, or to have the inventory database offline, or to have a
significant email outage. Yet despite these increasing needs and workloads, IT staffing hasn’t grown
nearly as much – and in many large data centers, overall headcount is down.

So how are large data centers dealing with greatly increased stuff to monitor and manage, and
higher service expectations with fewer people? It’s no secret; they do it with advanced systems

Copyright © January 2011 Gabriel Consulting Group, Inc.


management software allowing operators to view, monitor, and manage many more systems,
applications, and networks than ever before. With modern systems and application
management software, IT staffs can resolve potential service delivery problems well before users
begin to notice and ring IT service desk phones off the hook – thus avoiding any serious business
impact.

But what’s a small or midsized organization to do? These business owners face the same basic
challenges, albeit with far fewer resources. They too need to be responsive to an increasingly fickle
customer base that seeks immediate gratification to their product and service needs and wants via
any number of intelligent personal devices. SMBs (typically 50 - 1,000 employees) are now relying
on information technology just like their bigger business rivals. They’re hosting the same functions
and have seen a similar explosion in the number of applications and systems required to manage
vital corporate assets and lines of communication.

Where they dramatically differ, though, is in the amount of resources they can devote to IT
management. The vast majority of SMBs have small or nonexistent IT staffs, with maybe a handful
of employees who tend to be generalists. It’s not uncommon for the folks supporting enterprise
applications (ERP and CRM) to also be picking up the phone for desktop/mobile connectivity and
client applications issues. These same people are also in charge of data center hardware support –
including servers, storage, and network gear.

Midmarket IT shops tend to use a mishmash of point products to monitor or manage just a part of
the infrastructure. They end up jerry-rigging their own monitoring dashboards and cobbling them
together with their own custom code, or – worse yet – leveraging available ‘freeware’ that may be
both ill-fitting and under-featured. However, they don’t have a lot of time to screen-watch, so they
typically end up focused on whatever problem seems most critical – hardly the proactive and
efficient model required to succeed with such constrained resources.

The solution is a product that has the features and power of a full-fledged enterprise systems
management suite, but a learning curve and price point that matches SMB skill sets and resources.
The packages that large enterprises use have the features SMBs need, but are geared to more
sophisticated staffs who have deep and specialized knowledge. These solutions also carry price tags
that are out of the question for most SMBs.

IBM’s Tivoli brand team looked at this problem and saw an opportunity to develop a product set for
midmarket customers offering an almost turnkey solution: one that installs and adds value very
quickly. The result is the IBM Application Manager for Smart Business appliance, an application
monitoring and management solution along with IBM Service Manager for Smart Business, a
simplified, integrated IT service desk appliance.

Typically, when major software vendors like IBM attempt to fit their enterprise solutions to the
SMB market, they tend to cripple the product in some way; it still delivers some value, but not quite
enough value to keep customers from having to upgrade to their more expensive offerings as the
business grows. Another vendor approach is to preserve all the features of the enterprise product,
but lower the price so it’s in line with SMB budgets and limit its scalability with licensing terms.

Tivoli took the second approach: keeping the full feature set of the Tivoli enterprise products, but
limiting the Application Manager to 150 physical servers and 25 other network devices. IBM also

Copyright © January 2011 Gabriel Consulting Group, Inc. 2


made some significant changes to enhance the product's value and address the unique needs of the
SMB market.

The IBM Application Manager for Smart Business solution comes from an IBM Business Partner as
a pre-integrated hardware/software appliance that easily slides into any data center. It features a
simplified user interface, making it easy for even lesser-skilled IT employees to implement and
manage midmarket data centers. There’s even a robust set of user help screens and terminology
explanations. Remember, the Application Manager has all the functionality of the enterprise
monitoring solution; the only differences are the front-end and a few minor functions that really
aren’t relevant to small infrastructures.

IBM Application Manager for Smart Business Solution Components

We had the opportunity to see IBM demonstrate the Application Manager appliance installation
and operation in one of its labs. While we were there, we had access to a product expert, who talked
at length about use cases and customer benefits. He gave us the 'nuts and bolts' view of how the
product works, and apart from the application monitoring (which we’re going to discuss in detail),
two other components complete the solution.

1) IBM Lotus Foundations Start
Foundations is – well – the foundation of the Application Manager and other solutions. This
software layer acts as a gateway from the enterprise to the outside world and provides a firewall,
virus scanning for email and files, and a VPN mechanism. It has all of the services necessary to
support the enterprise including Apple/Windows file servers, a web server, DNS server, DHCP
server, LDAP directory server, IMAP mail server, POP3 email server, and others.

Administrators can always see system status and system utilization, plus the same for networks and
services. They can set alerts so that exceptions, such as a service failure or unusually high load, are
brought to attention of the administrator. Foundations also provides user administration
functionality – a single process to give users network access and permissions. It allows team- and
role-based security, and also sets up password and username policies.

The user management mechanism is much simpler than those we’ve seen with Windows and Linux
systems, and it makes adding users a quick and simple process. There are also plenty of help
screens that will walk the neophyte through almost every process. All of these features are
integrated and installed in a single operation, along with the Application Manager solution.

2) IBM System x server
IBM x3200 server hosts Foundations and the Application Manager software stack. One of the
potential reservations we had about the appliance approach is that it could become a bottleneck as
user online activity grows. We no longer see this as a problem after looking over the specifications
of the x3200 system. It is equipped with a single quad-core Nehalem EX processor and up to 32GB
of memory (though Application Manager only requires 4GBs). As importantly, it has plenty of
connectivity – multiple gigabit Ethernet ports – so that one or two ports can be directed to the
outside world while maintaining chip redundancy (no single chip failure can cripple the machine).
Application Manager can also run on third-party systems as long as they meet minimum
specifications.



Keeping an Eye on Systems

Installing Smart Business Foundations and the IBM Application Manager for Smart Business
solutions are simple processes that can be accomplished by just about anyone who has ever used a
personal computer (with the possible exception of my mother). Upon installation, the user sees the
welcome screen, also called the launch pad...

The launch pad is split into two sections, Operational and Administrative, which offer an easy-to-
understand, task-based approach. The first task is to have the software discover systems and bring
them under management by the Application Manager. The package will search the network and
find both systems and network-attached devices and return a list of what it found, as we can see on
the screen shot at the top of the next page.

The discovery process is full of surprises for many midmarket clients. Often they will suddenly find
network-attached devices that they either lost track of or didn’t know existed. Sometimes it’s old
hardware that used to run a service that’s no longer required. But the system is still there, taking up
space, using power and generating heat – and perhaps turning into a security hole as its software



goes un-patched. When these systems are found, IT shops often redeploy them in new roles,
allowing them to cross some projects off their list that were stalled due to funding or facilities
shortages.

As you can see, the
Application Manager
magically identifies
discovered assets by
host name and IP
address, and then shows
the operating system
and version. It also
shows how confident it
is in the discovery data.

It goes without saying
(but we’ll say it anyway)
that the data provided
by Application
Manager is much more
rich and accurate than
what exists in most
SMBs, some of which
track systems with
sporadically updated,
homemade
spreadsheets.

To the bottom left is a
help screen that explains the process and walks first-time users through it. Note that in addition to
x86 operating systems, we also see a filter for IBM’s AIX Unix operating system above the
discovery pane. Unlike Microsoft alternatives, both Foundations and the Application Manager
work with major Unix systems (Solaris, HP-UX, and AIX) as well as both Linux and Windows –
plus a huge number of network-attached devices.

The next step in the process is to click on the
“Deploy OS Agent” button to start managing the
systems that we discovered. This is a pretty quick,
one-time process, taking just a minute or two for
each item on the list. The agents are lightweight,
causing low overhead; however, if this is a
concern, agent-less monitoring can be used –
although the data received from monitoring will
be somewhat reduced, limited to SNMP data
alone.

Once they're operating, the OS agents track just
about everything you’d want to know about a
system – including CPU utilization, memory
utilization, disk utilization, which processes are



running, and a host of other stats in easy-to-understand graphs (previous page, bottom left corner).

Armed with this information, the IT staff can make intelligent decisions about balancing the
existing applications load across the available servers and determine which units have the potential
to host additional services without becoming overtaxed. Application Manager can be configured to
log the status of all components at user-defined intervals, and it will keep historical data for
reporting purposes.

The real key is that administrators can set it up so that it sends out alerts based on the condition of
any system attribute (or combination of attributes). There are a large number of situational alerts
preconfigured with the system (things like high CPU utilization, high memory utilization, etc.), but
users can easily set up their own alerts with their own thresholds.

A common complaint we frequently hear (usually in after-hours conversations) from midmarket IT
chiefs is, “We don’t know what the heck is attached to our network at any given time – much less
whether it’s performing or not.” Application Manager not only shows them exactly what’s on the
network and how it’s performing, but it also lets them easily see if a particular system can host
another service or workload. This means they can get more usage out of their existing systems and
can deploy new apps without having to buy more hardware.

Application Monitoring Made Easy

To manage applications, administrators simply deploy agents (or use the agent-less model) onto
the systems or virtual machines where the applications reside. Tivoli has a reasonably long list of
packages that it can monitor, beginning with Windows, Linux, and Unix operating systems.
Common SMB applications such as Microsoft Exchange and Lotus Domino email servers are
covered, as are Active Directory, BizTalk, Sharepoint, and other Microsoft platform software
offerings like Active Directory, Cluster Server, and Hyper-V virtualization technology. Supported
databases include Oracle, DB2, SQL Server, and Sybase solutions. Other applications can be
monitored with a customizable ‘universal agent’ that can be configured to track many application
attributes.

What’s monitored really depends on the application. For example with Exchange, Application
Manager will track average delivery time for both internal and external email, messages per
second, the number of connections, and even provide a list of current connections – along with
other important measures of email performance. Any of these characteristics can be used to set up
operator alerts, which can give admins a chance to correct performance or availability problems
before they impact users. No service disruptions means higher employee productivity – which is
how SMBs compete and win against larger competitors.

For databases, Application Manager tracks performance data such as read and write performance,
the percentage of session locks, row sorts, number of transactions, and a host of other
characteristics. This allows administrators to monitor database usage and performance – again,
heading off problems before they affect the business.

IBM Application Manager for Smart Business will also monitor applications inside virtual
machines including VMware ESX installations, giving operators vital information about the
resource usage of each VM and the system as a whole. The lack of this type of detailed intelligence
is what keeps many SMBs from adopting virtualization technology in the first place, and leads to
terribly underutilized assets in these days of multi-core processor technology.



Once agents are
deployed on network
devices, operating
systems, and
application instances,
the entire
infrastructure is
graphically depicted,
allowing operators to
see at a glance if
systems are offline
(denoted by a red 'x').
These types of bird’s-
eye views save
administrators lots of
time.

Administrators can also look
at a dashboard (at right)
where they can see system and
application status; again,
offline systems are shown in
red.

Both screens will allow users
to click on any managed
device and drill down into it to
see its status and all of the
other tracked statistics related
to it.



Dealing with Problems & Alerts

Perhaps the most important screen is the event console (next page). This is where operators will go
to see alerts and investigate problems or potential problems. With the right configuration of
alert thresholds, administrators should be able to use the events console as an early warning system
that will let them see situations they need to rectify before they turn into a calamity.

This is where we believe midmarket companies can save hours of time, and where they’ll recoup
much of the initial expense of the Application Manager appliance. In all the studies we’ve done,
labor is the most expensive element in any IT project. Easily consumable technologies like the
Application Manager that substitute 24x7 automation for eight hours per day of human eyes
should be on your company’s short list.

There are four panes on the event console. At the upper left is a navigation window that allows you
to isolate a particular system or application for further investigation. The window to the right is the
Situation Event Console , showing what’s specifically happening with the system in question. In
this screen shot there is a critical stoppage of a Linux process and a screen full of warnings on
other processes in the queue.



Below that pane is a mechanism allowing operators to acknowledge they’ve seen the event; this
shows others that the problem has been noticed and is being worked. This section is also used to
close or re-open a previously closed event. There is also a list of all of the open issues in the
infrastructure that gives a big-picture view of other problems enterprise-wide.

As we toured through the Application Manager, we were impressed by two factors in particular –
firstly, how easy it was to set up and use. In our opinion, anyone who has set up a server before will
feel instantly comfortable with the product’s simplified interface and become productive very
quickly. Secondly, highly technical administrators will revel in their newfound ability to see and
manage so much from a single pane of glass. They can easily see their entire IT infrastructure,
including system/application status and current/historical usage and utilization, and identify (and
fix) potential problems before they become critical.

These two observations led us to a new idea: this package may actually help turn novice system and
network admins into seasoned IT professionals. With IBM Application Manager for Smart Business
(and the integrated Foundations), they will finally be able to see how their network bandwidth is
being used. They will have enough data to be able to quickly decide which server has enough slack
capacity to host a new application. And they’ll finally be out of the vicious firefighting circle –
where they are so busy fixing the problems that blindside them daily that they can’t ever find time
to proactively address the conditions that are causing the problems in the first place. They may
even find themselves getting home in time for dinner more often – certainly a benefit, although
hard to quantify.

The system and application monitoring that the Application Manager provides is quite a bit more
rigorous and thorough than the vast majority of SMB administrators can provide. Couple this with
the alerts the package can generate, and it’s easy to see how this solution will save administrator’s
time and significantly improve overall IT performance, application availability, and infrastructure
resilience.

Does IBM Application Manager for Smart Business Pay Off?

It’s hard to put a dollar figure on increased IT performance and availability in the SMB. Every
business is unique and has different requirements and success drivers. It’s also difficult to measure
the financial impact of the Application Manager on IT labor. It obviously gives administrators
radically increased visibility into their infrastructure and operations. They can easily see current
and historic system loads, making it a simple task to figure out how to optimize utilization. The
solution will also alert them to changes in the IT environment so they can solve problems before
they become, well, problematic. But how much is that worth?

To get a feel for the economics of IBM’s Application Manager for Smart Business, we’re looking at it
in two different scenarios: 1. Application Manager vs. doing nothing and 2. Application Manager
vs. enterprise-level system management packages. First, let’s look at the cost of the appliance
solution and discuss how it compares to the status quo in most midmarket data centers.

On the cost front, there are acquisition costs for the Application Manager server appliance and the
initial software load, plus ongoing fees that cover technical support and updates. Licensing charges
for the monitoring software are based on flat, per server fees for any unit incorporating two
processor sockets (dual or quad core CPUs). In order to paint a realistic picture of the costs, we


put together a scenario where a midmarket customer needs to monitor 25 two-socket dual-core
servers and 25 two-socket quad-core systems. Here’s what the numbers look like:

Average Cost
Acquisition 3 Year Average per Server per
Cost Year 1 Year 2 Year 3 Total Cost Annual Cost Year
IBM System x3200 M3 hardware
appliance $3,200 -- -- -- $3,200 $1,067
Initial Lotus Foundations license $750 -- -- -- $750 $750
App Mgr license - 25 dual socket,
dual core servers $18,750 incl. $3,750 $3,750 $26,250 $8,750 $350
App Mgr license - 25 dual socket,
quad core servers $18,750 incl. $3,750 $3,750 $26,250 $8,750 $350

Total 3-year cost of ownership: $56,450 Average cost per year: $18,816

So for our test environment of 50 servers, the total cost of the Application Manager pencils out to
over $56,000 over three years--a very affordable price point adding only 10% more cost (per year)
to the initial hardware purchase. To really appreciate the application’s value, it’s important to look
at the costs in context, and compare them to the benefits delivered. Here are some points to
consider in your evaluation:

• Application Manager is comprehensive with components that monitor and manage systems,
applications, and network devices. Adopting the solution means you can rid yourself of numerous
point products used to manage bits and pieces of the infrastructure – which will definitely save
money.

• What if your most important app went down for a day or so? Compared to having little or no
formalized systems or apps management, it’s pretty easy to argue that using Application Manager
will increase application availability and decrease both planned and unplanned downtime. In our
example above, the average annual cost per server amounts to cheap insurance

• Application Manager enables greater efficiency in many ways. The most obvious example to us is
the way it can be used to implement and manage virtualization. Most SMBs have yet to adopt server
virtualization technology, mainly because they feel it’s a riskier and more complex environment.
Application Manager helps you see exactly which servers have room for additional workload, and it
will send out alerts when systems are over- (or under-) utilized. Average utilization on a non-
virtualized system is around 7%, and most customers see around a 5-to-1 virtual-to-physical server
ratio when they finally do adopt virtualization. This means you can reduce your physical boxes by a
factor of five and, over time, buy fewer new systems. There are considerable cost savings involved in
this – not just hardware costs, but also software licensing, along with power and cooling costs.

• Application Manager costs far less than hiring a new employee. With the large infrastructure in our
example, the software cost averages out to $19,000 per year – which is considerably less expensive
than the fully burdened cost of a new IT administrator. And Application Manager works 24 hours a
day, doesn’t forget or overlook any task it’s been assigned, and never calls in sick or takes a vacation.
Its advanced capabilities can also help make existing IT administrators more skilled and more
effective.


• The cost of Application Manager scales both up and down. If you use it on fewer servers, you pay
commensurately less. There isn’t a minimal fee or cover charge with this solution. The highest
benefit comes with covering the whole infrastructure, but there are certainly assets (like test servers)
that may not require sophisticated monitoring/management.

We obviously think that IBM’s Application Manager for Smart Business provides a lot of value for
SMBs and is certainly better than ‘manage by crisis’ or using a mishmash of non-integrated point
products to handle IT management and monitoring needs. But what about the alternative? Should
SMBs look farther up the food chain and adopt full-on enterprise IT management suites? Can they
get more value for their money and achieve even better results?

To answer these questions, we took at look at what enterprise versions of Tivoli would cost in our
50-server scenario above. We selected Tivoli because the features and functions lined up best with
Application Manager, since the solution is derived from Tivoli’s enterprise products. There are
other alternatives available from CA, HP, and others, of course, but we don’t believe that the
capabilities and costs for the enterprise versions of these products vary all that much from what
Tivoli is offering.

The Tivoli enterprise product closest to what most SMB customers would need is ITCAM for
Microsoft Applications. This package provides both agented and agent-less monitoring for
Microsoft operating systems, databases, email, and other enterprise services; however, it falls short
of Application Manager in quite a few areas:

• No coverage for Linux or Unix operating systems or applications.
• Does not monitor or manage non-Microsoft databases; no Oracle, DB2, or Sybase.
• More importantly, it doesn’t have any discovery mechanism to find applications or
operating systems on the network. This makes it more time-consuming to implement and
maintain, since managed components will need to be manually added to coverage.
• There is no network monitoring in this package.

Some of these holes can be filled in by adding another Tivoli product, ITCAM for Applications, at
additional cost. However, this still doesn’t give you much in the way of discovery or network
monitoring – which we believe are important capabilities, and very beneficial to SMBs. Keeping
that in mind, here’s how the numbers shake out:

Average Average Cost
Acquisition 3 Year Annual per Server
Cost Year 1 Year 2 Year 3 Total Cost Cost per Year
ITCAM for Microsoft Applications
license - 25 dual socket, dual core
servers $60,550 incl. $12,110 $12,110 $84,770 $28,257 $3,390
ITCAM for Microsoft Applications
license - 25 dual socket, quad core
servers $121,100 incl. $24,220 $24,220 $169,540 $56,513 $6,781

Total 3-year cost of ownership $254,310 vs. App Mgr $56,450
Average cost per year $84,770 vs. App Mgr $18,816

Adding ITCAM for Applications to add additional functionality adds only incrementally to the costs
of the above – anywhere from $7,000 to $13,300 on a three-year basis, depending on how many


applications need to be monitored on each server. This is a rounding error given the above
numbers, but it needs to be in there to provide a reasonably valid comparison to the Application
Manager. We also need to emphasize that the Application Manager price tag includes the server
on which it runs and full integration. Customers purchasing ITCAM for Microsoft and ITCAM for
Applications would need to provide host systems to run the packages – adding even more cost to
the enterprise solution.

One thing that we’ve learned from this exercise: enterprise versions of IT management suites are
pretty expensive vs. IBM Application Manager for Smart Business. This difference becomes a bit
more painful when you consider that the enterprise versions provide features and functionality that
don’t really apply to most SMBs (like monitoring mainframe applications and workflows) and don't
offer SMBs the things that are vital – like network monitoring and discovery. You can, of course,
get those functions, but they are licensed separately at additional cost.

The bottom line is that Application Manager provides more functionality than enterprise solutions
while costing 4-5X less.

The point here isn’t that enterprise management suites are bad or overly expensive – they aren’t.
It’s just that they are developed and engineered to handle a far greater variety of system
architectures, operating systems, and applications in far larger numbers and at the same time.
Large customers need these capabilities, but the vast majority of SMBs do not. This is why these
packages are bad deals for SMBs – they pay for a lot of features they will never use. On the other
hand, there aren’t any comprehensive system/application management packages that have the
capabilities of Application manager yet the simplicity of an appliance. It’s this open niche in the
market that caught IBM’s attention.

Summary & Recommendations

We like what we see in the IBM Application Manager for Smart Business and the IBM Tivoli SMB
strategy. As we mentioned above, we see the solution as filling an empty niche in the market – a
very full-featured and comprehensive enterprise IT management solution delivered in a form that
fits SMBs' skill levels and budgets. With Application Manager, we believe clients will see
immediate benefits in terms of service availability and greater IT utilization. We would recommend
that clients arrange for a demonstration and get more details on how it might fit into their own data
centers.

Entire contents © 2001- 2010 Gabriel Consulting Group, Inc. All rights reserved. This document may not be reproduced or
transmitted in any form by any means without prior written permission from the publisher. All trademarks and registered
trademarks of the products and corporations mentioned are the property of the respective holders. The information contained in this
publication has been obtained from sources believed to be reliable. Gabriel Consulting Group does not warrant the completeness,
accuracy, or adequacy of this report and bears no liability for errors, omissions, inadequacies, or interpretations of the information
contained herein. Opinions reflect the judgment of Gabriel Consulting Group at the time of publication and are subject to change
without notice.

phone / 503.372.9389
gcginfo@gabrielconsultinggroup.com
www.gabrielconsultinggroup.com


Gabriel consulting whitepaper on enterprise IT for SMB

More Related Content

What's hot (19)

Similar to Gabriel consulting whitepaper on enterprise IT for SMB (20)

More from Jyothi Satyanathan (19)

Gabriel consulting whitepaper on enterprise IT for SMB