Appliances and ROI

Overall I think it should be cheaper and in a customer's best
interest  to use appliances.  It gets to the issue of where knowledge
can be applied most effectively.   Of course this assumes that the

    a)  solve the problems of the market
    b)  are partitioned (in function or pricing model) in a way that makes them a reasonable fit
    c)  have a roadmap that address challenges as (or before) they become important to customers
    d)  can interact with other appliances and non-appliances in standardized ways to customize overall capability

appliance is like a contract between a customer and a vendor and
implies a level of trust that the vendor will embed some part of the
skill set of a highly trained specialist by incorporating lifecycle
best practices (structure, operations, performance, security, etc.) in
the appliance therefore making the benefit achievable by someone
without the specialized skills.

The initial outlay for an
appliance may be more expensive than the constituent elements, however,
if an appliance is well crafted and is a good fit for the customer,
then ROI should be achieved in a reasonable amount of time.  When
comparing the ROI of an appliance and a bespoke system, you should
consider not only the benefits, but the cost avoidance that the
appliance enables.

Areas to look for ROI include:

Installation/Setup time
- an appliance should be much more efficient to set up than the
equivalent non-appliance.  There are several ways that a good appliance
can streamline this phase including standardizing and documenting
interfaces with the environment, finding the right balance of packaged
vs configured at install time, enabling hands-free installs, etc.  Not
only does a good setup process reduce the labor costs when compared to
a bespoke system, it also reduces the time to get the system
productive, thereby enabling it to start generating revenue earlier.

- an appliance has the potential to embody peformance and capacity
planning knowledge from a broad range of environments.  If it is
designed with peformance in mind, it should be able to achieve strong
performance numbers compared to a bespoke system.  Higher throughput
enables the system to generate more revenue, and pushes off the need to
add capacity.

Day-to-Day Operations - an appliance
should require less skill to operate than a bespoke system since
operational best practices are included in the appliance.  This such as
log maintenance, state change, provisioning users/roles, etc. should
all be included in the appliance.  It should not be necessary to write
scripts around the constituent elements to glue them together. 
Specialized knowledge can be applied elsewhere, making the
operational/administrative cost of the appliance less than the
equivalent bespoke system.

Support and Maintenance -
support of an appliance should turn out to be much less expensive than
an equivalent bespoke system since the elements and configuration of
the appliance will be well understood by the vendor.  Not only will
this reduce downtime, but when there is a problem, it will reduce the
time required to analyze and fix the problem. 

with any contract, understanding the fine print is important.   An
appliance has to be designed to address these things, has to be priced
competitively, and has to be matched to the right type of problem.  If
it is, it should be a much more compelling option than building your


Writing from experience as someone who built several appliances, the cheapest way and the biggest return on investment is building one's own appliances.

Appliances built internally to the company will:

- be optimized for the task at hand
- be built on top of generic software and hardware components
- be extensible, well documented, and well understood
- be built on the LEGO(R) principle
- be capable of being versatile, as opposed to being optimized for one and one thing alone.

Perhaps the biggest piece if the "be built on top of generic software and hardware components", which ensures easy servicing and modularity.

Such appliances are optimized for company's internal processes, and the software is custom tailored to the organization, yielding 100% requirements compliance.

However, these long term gains are offset by short term costs: engineering of a software platform generic enough to be modular and optimized at the same time is costly and an incremental improvement process, which is offset by long term benefits.

The reason it is offset by long term benefits is that incremental improvement is very cheap and yields very good results with minimum amount of effort, if performed properly.

The added benefit is using off-of-the-shelf parts, which makes the whole unit become a FRU ("Field Replacable Unit") because of mass production and ultra-low cost.

In other words, a company which can afford an internal system engineering department will benefit tremendous long term financial savings... and spells doom for companies which make their profit off of selling appliances.

Posted by UX-admin on February 16, 2010 at 01:25 AM PST #

Post a Comment:
Comments are closed for this entry.

This is a space for me to post things I am thinking about. Most content will be related to datacenter operations and automation.


Top Tags
« July 2016