Detailed evaluation of various digital asset management (DAM) hosting service models - such as infrastructure as a service (IaaS) and software as a service (SaaS) - including costs, advantages and disadvantages, a set of prospective questions for DAM vendors and a use case example
The Work Ahead in Intelligent Automation: Coping with Complexity in a Post-Pa...
Key Factors to Consider When Hosting DAM in the Cloud
1. Key Factors to Consider
When Hosting DAM in the Cloud
The pros and cons of various service delivery methods, which will
enable organizations to choose an optimal digital asset management
hosting solution.
Executive Summary
This white paper provides a description and over-
view of various digital asset management (DAM)
hosting solutions, including cloud-based, on-
premises and hybrid structures, as well as the
advantages and disadvantages of each solution.
Our overall objective is to raise awareness and
help organizations evaluate various DAM host-
ing solutions, their costs and the IT resources
required to make more educated decisions about
how to best deploy a DAM system.
In the course of this paper we will also look at
a use case example and the associated costs of
different hosting models. Also provided is a scor-
ing model that can be used to compare different
hosting scenarios based on your organization’s
requirements, and a list of questions that can
be posed to prospective DAM hosting vendors to
assist with vendor selection and analysis.
DAM Service Models Defined
There are a variety of DAM service models; for
the scope of this paper we will focus on five varia-
tions which we believe are most applicable to an
organization seeking a DAM hosting solution. The
models we will evaluate include:
• On premises/hosted co-location data center.
• Infrastructure as a service (IaaS).
• Software as a service (SaaS).
• IaaS and on-premises hybrid.
• SaaS and on-premises hybrid.
On Premises/Hosted Co-Location Data Center
Systems that physically reside on premises are
typically hosted in a data center or server room
that is owned and operated by the end-user orga-
nization. For the purpose of this paper we will
also include the use of a co-location data center
as another form of on-premises solution, the pri-
mary difference being that the co-location data
center facility is owned and operated by a third
party. However, the system hardware and soft-
ware are still owned and operated by the customer.
• Cognizant 20-20 Insights
cognizant 20-20 insights | november 2014
2. 2cognizant 20-20 insights
The on-premises solution illustrated in Figure
1 is the most traditional model. In this case all
infrastructure, servers, storage, networking and
applications are managed and maintained by the
customer’s IT organization. In this scenario, a
DAM application would be installed on in-house
hardware and would leverage infrastructure and
systems owned and managed by the company’s
internal IT team.
Figure 1 is broken up into a set of six different
resource groups that are defined as follows:
• DAM systems: This represents all of the
servers that make up the DAM application
as a whole, and these could range anywhere
from a single server to dozens of systems,
depending on the business requirements and
the DAM solution selected. The DAM system
includes the core DAM application, Web
servers, database systems, search and indexing
services, annotation, file transformation/
transcoding and file delivery systems, among
others. Anything that would be considered
a component of the DAM system as a whole
would be represented here. Management of
these servers is divided into the following
layers:
>> Hardware/virtualization management (VM).
>> Operating system (OS) management.
>> DAM applications management.
>> DAM administration.
In the on-premises model, all of these layers
would be managed by the customer’s IT team.
Note that for the scope of this paper, DAM appli-
cation management refers to management
and engineering of the underlying DAM appli-
cations (systems administration, development,
database administration, etc.) whereas DAM
administration refers to user administration
and light configuration of the DAM applications
(creating users, granting access, managing
workflows, etc.).
• Network: This includes any and all networking
equipment such as network circuits, Internet
access, routers, switches, load balancers,
firewalls, cabling and associated network
DAM System
Ancillary Services
Customer
Managed
Network
Storage
Hardware/VM Hardware
Admin
Hardware
Admin
Hardware
Admin
OS
DAM App Mgmt.
DAM Admin
Directory Services
DNS
SMTP
Monitoring
Management Tools
Third-Party Services
Agencies/Distributors Company Site/Workgroups
Archive/Backup
SOA/ESB
Editing Systems
Publishing Systems
Rights Mgmt.
Order Processing
Hardware/VM
OS
Archive Apps
INTERNET
DAM USERS
PRIVATE CIRCUIT
OR LOCAL CONNECTION
Hardware
Admin
In-House DAM
Figure 1
3. cognizant 20-20 insights 3
management systems. Note: This is divided
into hardware and administration management.
• Storage: This refers to any and all storage
whether it is a storage area network (SAN),
network attached storage (NAS) or locally
attached storage. This is divided into hardware
and administration management. In this case,
storage administration refers to fairly light-
weight tasks — such as storage allocation and
storage performance configuration — that
could be done from a power-user perspective.
• Archive/backup: This is the archive/backup
system which is typically shared within the
enterprise. This example illustrates a backup
server and an offline storage system (tape or
other). Management of the system is divided
up into hardware, OS and the archive/backup
application.
• Third-party services: This consists of
additional shared services typically owned and
managed by your organization or a business
partner that is not part of the DAM environ-
ment but which exchanges information with the
DAM systems. This can include services such as
an enterprise service bus, single sign-on, third-
party editing and publishing systems, author
submission systems, rights management, order
management and billing systems.
• Ancillary services: This includes services that
are necessary for the DAM system to operate
and to be managed properly but are not part
of the DAM system itself. These are usually
low-level shared infrastructure services such
as directory services, domain name service
(DNS), simple mail transport protocol (SMTP,
aka e-mail), system monitoring and system
management tools.
On-Premises Advantages
• High degree of customization and integra-
tion: An on-premises solution provides the
greatest degree of customization over the
DAM application stack and the underlying
physical infrastructure. This can be of signifi-
cant benefit in an environment that has a high
degree of integration with third-party systems
and/or requires significant application integra-
tion and development.
• Leverage existing infrastructure: On-premis-
es DAM solutions can take advantage of shared
servers, storage, networking and related infra-
structure where excess capacity exists and the
technology is feasible. This has the potential
to offset some capital and management costs.
Management processes and procedures such
as change control, data backup procedures and
enterprise monitoring can also be more readily
integrated with an on-premises DAM solution.
• Control over the environment: An on-prem-
ises solution delivers the greatest level of
control over the entire environment: servers,
networking, data storage, security and service
levels. This may be a requirement in certain
regulatory compliance scenarios so it is
important to identify any compliance require-
ments as part of the decision-making process
when selecting a DAM hosting model.
• High-speed connection for local users and
systems: Organizational users and on-prem-
ises DAM systems are usually connected via
high-speed LAN connections (1 to 10 gigabits/
sec.). This is ideal for work-in-progress
scenarios where large files or video content
needs to be manipulated or edited. This is
also an advantage where large files need to be
moved about between systems within a facility.
High-speed LAN connectivity can be procured
relatively inexpensively when compared with
high-speed WAN or Internet circuits.
On-Premises Disadvantages
• Significant capital investment: Large cash
outlays are usually required for any on-prem-
ises solution. This includes the procurement
of servers, software, storage, networking
equipment, rack space, power, cabling and a host
of other costs. In the best-case scenario, some
of these costs can be reduced if your organiza-
tion makes use of virtualization technologies
and can leverage existing infrastructure.
• Higher IT administrative costs: There are
resource costs associated with the routine
maintenance and management of IT infrastruc-
ture. This includes time spent by system admin-
istrators, network administrators, storage
administrators, DBAs, application administra-
tors, security specialists and IT management.
The time and resource costs dedicated to
managing and troubleshooting these systems
is usually very significant, especially when
evaluated over the long run.
• Time-to-market/deployment: The time
required to deploy a new DAM solution can
be significant, depending on the status of
your organization’s infrastructure. Organiza-
tions with more modern infrastructure that
4. cognizant 20-20 insights 4
leverage virtualization, and that have ample
infrastructure to spare, will have a much easier
time deploying a DAM system than a more tra-
ditional on-premises environment that needs
to procure, install and configure equipment.
The process to procure and physically install
hardware alone can range anywhere from
two weeks to two months depending upon
the size of the implementation and your ability
to execute.
• Limited scalability/flexibility: On-premises
data centers typically have limited scalability
capabilities beyond which they require a sub-
stantial amount of new capital investments to
grow the environment. This could be servers,
storage, networking bandwidth, physical space,
power, etc. Also, additional capital investments
beyond what is currently required are typically
made to provide future room for growth. This is
an inefficient use of capital. For example, if your
organization needs 10TB of storage, it won’t
purchase exactly 10TB of storage, but 15TB or
maybe 20TB to account for future growth. The
extra money spent today for future growth is
not utilized until some time in the future at
which point the storage is also less expensive;
the extra money spent today could have
been spent on something more immediately
necessary.
• Management of the application stack: This
could be considered either a pro or a con
depending upon the degree of control and
flexibility your organization wants over its
DAM applications. The disadvantage here is
the additional management time and cost
required by your IT organization to administer
and maintain the software applications that
reside on-premises (especially when compared
to an SaaS model). This can be a substantial
cost in terms of IT resources and depends upon
the size and complexity of the DAM solution
implemented — and the capabilities of your IT
organization. DAM systems are not common
commodity applications, and unfortunately,
there are very few trained and certified DAM
administrators.
Infrastructure as a Service
Infrastructure as a service (IaaS) is a cloud-based
model whereby all of the infrastructure services
are managed and maintained by the IaaS cloud
provider: servers, storage, networking, etc. The
DAM System
Ancillary Services
Customer
Managed
Cloud Provider
Managed
Network
Storage
Hardware
Admin
Hardware
Admin
Hardware
Admin
OS
DAM App Mgmt.
DAM Admin
Directory Services
DNS
SMTP
Monitoring
Management Tools
Third-Party Services
Agencies/Distributors Company Site/Workgroups
Archive/Backup
SOA/ESB
Editing Systems
Publishing Systems
Rights Mgmt.
Order Processing
INTERNET
DAM USERS
PRIVATE CONNECTION
HW/VM
HW/VM
HW/VM
HW/VM
HW/VM
HW/VM
HW/VM
HW/VM
HW/VM
HW/VM
Hardware/VM
Hardware/VM
Archive Apps
Hardware
Admin
Infrastructure as a Service
Figure 2
5. cognizant 20-20 insights 5
primary distinction here is that the IaaS provid-
er does not manage the actual DAM applications,
or any applications for that matter including the
OS. In the IaaS model, the customer is respon-
sible for managing and maintaining the OS and
DAM-related applications. Some common IaaS
providers include Amazon AWS, Microsoft Azure,
IBM Cloud and RackSpace.
Figure 2 illustrates how hardware and core infra-
structure services are managed by the IaaS
provider (in green); the blue sections are typically
the management responsibility of the customer.
Some level of administration over the infrastruc-
ture is usually handled by the customer using
abstract tools provided by the IaaS vendor, which
simplifies what would otherwise be more complex
infrastructure administration. In an IaaS model,
the management of these services is divided up
as follows:
• DAM systems: Keep in mind that in an IaaS
cloud,alloftheserversarevirtualizedfromyour
organization’s perspective. In the IaaS model,
management of the server hardware and virtu-
alization/hypervisor layer is performed by the
IaaS provider. The customer is responsible for
managing the OS and all of the DAM-related
applications on the servers.
• Network: The underlying physical network
components are managed by the IaaS
provider; this includes Internet access, routers,
switches, load balancers, firewalls, cabling and
associated network management systems. The
key difference here is that the administration
of virtualized network devices is handled by the
customer, and includes tasks such as defining
firewall policies, IP addressing, subnetting,
network routing, etc. Network administration
of IaaS components is performed at an abstrac-
tion layer that does not typically require a fully
qualified network administrator to manage.
IaaS providers also typically offer direct
network connections between the customer’s
facilities and the IaaS cloud for additional per-
formance and security (if so desired).
• Storage: This includes a variety of different
storage including network attached, local and
object storage. The underlying hardware infra-
structure is managed by the IaaS provider and
administrative functions such as disk allocation,
performance configuration and file system
configuration are managed by the customer
via the abstracted and simplified storage tools.
• Archive/backup: In an IaaS model, the archive
system hardware, including the archive server/
system(s) and high-capacity storage devices
are managed by the IaaS provider behind the
scenes. The customer is typically responsible
for managing the archiving and backup appli-
cations, schedules and related operational
processes.
• Third-party services: This consists of
additional shared services typically owned
and managed by the customer or a business
partner that is not part of the DAM system,
but which interfaces with the DAM environ-
ment to exchange information. This can
include services such as an enterprise service
bus, single sign-on, third-party editing and
publishing systems, author submission systems,
rights management, order management and
billing systems. It is rare to find organiza-
tions that have migrated all of their third-
party services to an IaaS cloud. What is more
common is a hybrid configuration where some
or all third-party services are hosted on-prem-
ises or elsewhere.
• Ancillary services: Most ancillary services are
provided and managed by the IaaS provider.
This includes services that are necessary for
the DAM system to operate but are not part of
the DAM system itself. These are usually low
level shared infrastructure services including
directory services, DNS, SMTP (e-mail),
monitoring and management tools. Third-party
services and other ancillary services can also
be hosted within the cloud or on-premises as
a hybrid model.
IaaS Advantages
• Minimal capital investment: With IaaS, the
customer does not need to purchase hardware
infrastructure up front, and only pays for the
resources needed at the moment. This is ideal
for organizations that prefer an operating
expenditure model over a capital one.
• Minimal infrastructure management: The
underlying core hardware infrastructure of
an IaaS offering (servers, storage, networking,
etc.) is managed and maintained by the IaaS
provider. High-level administrative control
over these components is provided but in an
abstracted and simplified way that eliminates
management of the underlying core systems.
So instead of requiring a network, storage and/
or systems engineer, the customer is more
likely to make do with a single systems admin-
istrator or power user.
6. cognizant 20-20 insights 6
• Pay-as-you-go: This model optimizes the use
of capital because the customer only pays for
resources consumed. In a traditional on-prem-
ises model, it is not uncommon for organiza-
tions to spend an extra 25% to 50% on capital
equipment to provide “scalability” in the
coming years. This is tied up money that could
otherwise be put to immediate use.
• Time-to-market/deployment: The time it
takes to procure the infrastructure required
to build a DAM solution using an IaaS model is
incredibly short when compared to a tradition-
al on-premises procurement and build cycle. In
general, it is safe to say that procurement of
infrastructure resources in an IaaS model can
be done within a matter of hours, because it
is essentially all virtualized, compared with an
on-premises solution that can take weeks to
months to procure and physically install into a
data center.
• Strong scalability/flexibility: The ability to
grow storage, servers, networking and other
infrastructure services within an IaaS cloud is
due to the high degree of virtualization these
environments provide. There is an incredible
degree of scalability in an IaaS cloud, which
allows you to grow and shrink DAM needs
dynamically. This flexibility allows your orga-
nization to easily change resource alloca-
tions to fit real-time needs. Resources can be
reduced if demand declines, thus saving costs,
or resources can be increased very quickly
if an unexpected demand arises. Addition-
ally, larger IaaS providers have the ability to
distribute DAM services to regions throughout
the world. This can be beneficial in situations
where your organization wants different levels
of DAM content or services to be placed locally
to a given community of users for performance
reasons or per legal requirements.
• High degree of customization over the appli-
cation stack: In an IaaS model, the customer
has complete autonomy over the DAM applica-
tions stack, including everything from the OS
on up. This provides the greatest degree of cus-
tomization and control over the applications,
and facilitates the integration of third-party
software components.
• Security: This is a controversial subject, and
it is commonly cited as a significant concern
regarding the use of a third party to host and/or
managecustomercontent.Acloseevaluationof
enterprise-class IaaS providers demonstrates
a superior level of security when compared to
traditional on-premises or SaaS-based DAM
solutions. Bear in mind a primary tenet of any
major IaaS provider is security, which is applied
in layers, from physical separation of duties all
the way up the stack to software-based access
control lists and schemes. Very few customer-
hosted data centers have the ability to provide
the same degree of security available from an
enterprise-class IaaS provider. With this said,
also bear in mind that not all IaaS providers are
created equal. The level of security will vary
among vendors, so it is important that your
organization carefully evaluates the security
methodology used by the IaaS provider.
• Cost-effective archive/backup and disaster
recovery: Data backup solutions can be less
costly using IaaS over a traditional on-premises
solution. When cost for IT staff to manage and
administer any backup solution (servers, robots,
tapes, software, etc.) is included, organizations
typically see a significant administrative time
and cost savings when using an IaaS provider.
In an IaaS model, control over the operations,
schedule and software is maintained by your
IT organization, but the underlying complexity
and management of the system is abstracted
away. IaaS providers also typically have a large
number of options when it comes to disaster
recovery and business continuity needs. This
includes customer selection of physical regions
where systems and content is stored, replica-
tion and various levels of redundancy that can
be available at a very modest cost.
IaaS Disadvantages
• Vendor lock-in: Many IaaS vendors have a
proprietary hypervisor and/or proprietary
virtual machine images. This makes it difficult
to migrate virtual machines and systems
to another vendor or in house. Some IaaS
vendors promote the use of OpenStack and
other standard virtual machines such as
VMware to make migration easier. Additionally,
any custom developed applications that make
use of a given IaaS API set will have to be rede-
veloped if moved to another IaaS platform or
brought in house. The cost to extract a large
repository of data out of the IaaS provider can
also be substantial, as there are fees normally
associated with egress data which can facilitate
the lock-in effect.
• Limited SLAs: Service level commitments by
IaaS providers may not meet your organiza-
tion’s requirements. Historically, SLAs from
IaaS providers have not been as encompassing
7. cognizant 20-20 insights 7
as those developed by traditional data center/
co-location hosting providers. It is important
that your organization carefully reviews
what SLAs are offered by the IaaS provider,
especially around support, maintenance
windows, response time, mean time to repair
and, especially service credits.
• Expenses can grow unexpectedly: In an
expense model, you pay for services as you
use them. It is very important that these costs
are closely monitored, as they can grow very
rapidly and unexpectedly. Unlike a Cap-Ex
model where most of the costs are pre-allocat-
ed, your organization can experience unpre-
dictable monthly expenses based on the utili-
zation of the DAM system. This can be caused
by a spike in repository storage needs, file
deliveries, CPU processing (i.e., transcoding),
etc. It is highly recommended that if your orga-
nization decides to utilize an IaaS provider, be
sure to implement a rigorous process of cost
monitoring and review to reduce the risk of
unexpected costs.
• Tightly coupled DAM applications: Tightly
coupled DAM applications are integrated with
other systems and services, typically making
direct API calls to a given server, service or
device that does not make use of a queuing,
service bus or load balancing architecture.
DAM applications that are tightly coupled
cannot take advantage of IaaS scalability
features; the only scaling options are to replace
virtual servers with bigger ones — which is not
as effective as adding more virtual servers to
a farm. To this point, it is important that you
clearly understand how the DAM applications
scale: vertically (bigger servers) or horizontally
(more servers), which is the preferred method.
• Network performance and cost: With an
IaaS provider your organization usually has
the option to leverage an Internet connection
and/or a direct/private WAN connection. In
situations with long-form content (such as
video) or a significant amount of work-in-prog-
ress material, you may require a dedicated
high-speed connection between end users and
the IaaS cloud provider. The cost to procure
high-speed dedicated connections is relatively
high, and there are usually bandwidth charges
associated with content downloaded from the
provider. When these additional bandwidth
fees are assessed, your organization may find
the IaaS solution more costly in the long run
than a more traditional on-premises solution
where network (and storage) traffic can be kept
local to the users and systems.
• Customer manages the application stack:
This could be considered either a pro or a
con, depending upon the degree of control
and flexibility you want over the DAM applica-
tions. The disadvantage here is the additional
management time and cost required by your
IT organization to administer and maintain
the software applications that reside within
the IaaS cloud. This can be a substantial cost
in terms of IT resources, and it depends upon
the size and complexity of the DAM solution
implemented and the capabilities of your IT
organization.
Software as a Service
In a SaaS model, all underlying infrastructure
services as well as management of the DAM
application are managed by the SaaS provid-
er. From a management perspective, this is the
simplest and usually the least expensive scenar-
io. Figure 3 (next page) illustrates a typical SaaS
solution for a DAM system.
Figure 3 shows that management of the entire
ecosystem (with the exception of some DAM-
related administrative functions) is owned and
managed by the SaaS provider. Key differenc-
es between the SaaS, IaaS and on-premises
models include:
• Many DAM SaaS providers actually leverage an
IaaS provider behind the scenes; however, the
SaaS provider is still ultimately responsible for
managing the overall service to the customer.
• The DAM SaaS provider is not always the DAM
vendor that developed and sells the product.
Some DAM vendors leverage business partners
to provide and manage a SaaS implementation,
using the vendor’s DAM product.
• Third-party services are absent in a SaaS model,
but rather are hosted by your organization or a
trusted partner.
• Direct network connections between your
organization’s facilities/data center and the
SaaS provider are not typically available.
SaaS Advantages
• Minimal capital investment: As with the IaaS
model, there is no up-front capital investment
required to use a SaaS DAM. There may be an
up-front initiation fee, and there is typically
8. cognizant 20-20 insights 8
Software as a Service
a monthly recurring subscription fee based
on user licenses and other features that are
subscribed to.
• Pay-as-you-go: In a SaaS model, your orga-
nization will typically pay a monthly subscrip-
tion fee that should be adjustable based on
user count and resource usage (transactions,
storage, compute, bandwidth, etc.). However,
these models vary considerably by SaaS
vendors. It is important that your organization
obtain a clear understanding of the subscrip-
tion model costs and all contractual terms
of the agreement with the SaaS provider to
avoid unexpected fees that could occur as your
requirements grow.
• Minimal IT administrative cost: The SaaS
model entails by far the lowest administra-
tive costs of all the models highlighted. In
this scenario the only responsibilities of the
customer’s IT organization are usually light
administration and configuration of the DAM
application such as creating users, granting
access, managing workflows and other light-
weight administration tasks.
• Time-to-market/deployment: The time it
takes to implement a SaaS solution is faster
than any of the other models presented herein.
The amount of time involved onboarding a new
customer will vary from vendor to vendor, and
will depend upon complexity and feature set,
but overall a SaaS model is typically the fastest
way to get a DAM service up and running.
• Software upgrades and maintenance: A
benefit of using a SaaS model is that upgrades
and maintenance are all handled by the vendor,
and are usually included with the subscrip-
tion fee. This also ensures that your organiza-
tion has the latest product features available
without having to go through an upgrade
process yourself. Ideally, this is also transpar-
ent to your users.
DAM System
Ancillary Services
Network
Hardware/VM Hardware
Admin
Hardware
Admin
Hardware
Admin
OS
DAM App Mgmt.
DAM Admin
Directory Services
DNS
SMTP
Monitoring
Management Tools
Third-Party Services
Agencies/Distributors Company Site/Workgroups
SOA/ESB
Editing Systems
Publishing Systems
Rights Mgmt.
Order Processing
INTERNET
DAM USERS
Customer
Managed
SaaS Vendor
Managed
SOA/ESB
Editing Systems
Publishing Systems
Rights Mgmt.
Order Processing
UNAVAILABLE
Storage
Hardware
Admin
Archive/Backup
Hardware/VM
OS
Archive Apps
Figure 3
9. cognizant 20-20 insights 9
SaaS Disadvantages
• Limited degree of customization: Typically,
most SaaS services have limitations on how
customized the solution can be. In part, this
is due to the fact that many SaaS installations
are multi-tenant and have some shared depen-
dencies, or from a management perspective it
is easier for the SaaS provider to limit custom-
ization of the DAM system. This can be a sig-
nificant hurdle for enterprise customers that
usually require a significant degree of custom-
ization. Some SaaS vendors only offer a subset
of the API services available compared with
their on-premises version. It is important that
your organization understands the differences
in functionality between the vendor’s on-prem-
ises and SaaS versions of the product.
• Challenges integrating with third-party
services: Some DAM SaaS products may have
limited third-party integration capabilities
when compared with on-premises solutions.
Ideally, the DAM vendor will provide the same
level of integration for both service delivery
models, but this typically varies among DAM
products, and needs to be validated with the
vendor. Also, because third-party services
do not reside within the SaaS infrastructure,
there are additional security complexities and
performance considerations that can arise
from having to communicate with third-party
services via the public Internet. Additionally,
customers with legacy third-party systems that
do not support Web services (used to interface
with Internet-enabled services) will not be able
to easily integrate with a SaaS DAM without
having to rewrite their legacy applications.
• Vendor lock-In: Migrating from a given SaaS
provider may prove difficult, since exporting
the customer repository, database configu-
rations and related data are not typically
available from a SaaS provider. This needs to
be clarified with the provider up front so your
organization understands the options should
the need arise to migrate to a different DAM
solution or platform.
• Infrastructure management is not a core
competency: The primary business of a
DAM vendor is the development and sale of
its products and services; its business is not
the management of infrastructure (servers,
storage, networking, data backups, security,
etc.). This is not to say that DAM SaaS vendors
do a poor job managing infrastructure services.
Rather, this is to say that infrastructure is
not a core competency of a typical DAM
SaaS vendor, and as such they are unable to
provide the same level of infrastructure capa-
bilities and support that an IaaS vendor, on-
premises/co-location provider or dedicated IT
infrastructure staff typically can. The level of
capability varies widely among SaaS vendors.
The best-case scenario would be a SaaS
provider that leverages a major IaaS provider.
This gives your organization the best of both
worlds. Again, bear in mind the SaaS provider
is still ultimately responsible for the overall
management and delivery of the entire DAM
service even if it leverages an IaaS vendor to
provide infrastructure services.
• Network performance limited to public
Internet: Typically in a DAM SaaS model, all
means of communication is across the public
Internet. Direct/private WAN connections
between a customer site and the SaaS cloud
provider are not usually
supported. This puts limi-
tations on the quality and
availability of bandwidth
to transfer content. This
can be a problem where
large-form content (such
as video) or a signifi-
cant amount of work-in-
progress content is being
supported. It is recom-
mended to ask the SaaS
vendor during the initial
evaluation if it supports
direct network connections
as a potential means to
address bandwidth issues
should the need arise. It is also important to
note that the cost of a dedicated/direct circuit
can be relatively high.
• Limited archive/backup and disaster
recovery: SaaS providers do not typically
provide archiving, backup or disaster recovery
services for individual customers, and in
those cases where they do, the services tend
to be limited. On the whole, SaaS vendors
do back-up data stored within their systems.
However, because of shared tenancy and other
factors, this is a difficult service for many of
these companies to provide at a granular level
for the individual customer. When evaluating a
SaaS vendor, clarify what archive and backup
services it can provide and request specific
details about how and when it would recover
your organization’s data and restore the DAM
SaaS service in case of failure.
Customers with
legacy third-party
systems that
do not support
Web services will
not be able to
easily integrate
with an SaaS DAM
without having to
rewrite their legacy
applications.
10. cognizant 20-20 insights 10
• Limited SLAs: Service level commitments by
SaaS providers may not meet your require-
ments. SLAs of SaaS providers will focus more
on support of the application, and are not as
encompassing as those provided by traditional
data center hosting providers when addressing
the underlying infrastructure services. You
need to review what SLAs are offered by the
SaaS provider, especially around support,
maintenance windows, response time, mean
time to repair and service credits.
IaaS and On-Premises Hybrid
A very common model is some combination of the
IaaS and on-premises models wherein the core
DAM services reside in the IaaS cloud and sup-
porting services reside on premises or at another
location. There can be numerous combinations of
these services and models. For simplicity, Figure 4
illustrates a common model in which DAM servic-
es reside within the IaaS cloud, and the ancillary
and third-party services reside on premises.
In the IaaS and on-premises hybrid model, your
organization experiences all the advantages and
disadvantages that were previously described for
the IaaS model. The primary difference is that
your organization has the ability to leverage pre-
existing enterprise services that already reside
on premises. This includes ancillary services
such as directory services, single sign-on, enter-
prise monitoring, reporting tools and third-party
services such as an enterprise service bus, edit-
ing and publishing systems, rights management,
order processing and any other service required
to communicate with your organization’s DAM
system. Web service APIs would typically be used
for communication among the environments.
Another feature of this hybrid model is the capa-
bility to archive/backup/replicate system data in
both the IaaS cloud and on premises. This may
be of particular interest for customers who desire
a local copy of their DAM repository and related
data for security, disaster recovery and other
purposes.
There is also a great deal of network flexibility
available with this hybrid model, since it supports
direct network connections among the IaaS
cloud, on-premises data center(s) and customer
sites. Direct connections are optional, but if per-
formance or security is a concern now or in the
future these options are generally available.
DAM System
Ancillary Services
Ancillary Services
On-Premises Data Center
Customer
Managed
Cloud Provider
Managed
Network
Hardware
Admin
Hardware
Admin
Hardware
Admin
OS
DAM App Mgmt.
DAM Admin
DNS
SMTP
Agencies/Distributors Company Site/Workgroups
Archive/Backup
INTERNET
DAM USERS
PRIVATE CONNECTION
PRIVATE CIRCUIT OR LOCAL CONNECTION
PRIVATE CONNECTION
Hardware/VM
System Monitoring
Directory Services
Enterprise Monitoring
Enterprise Mgmt. Tools
Third-Party Services
SOA/ESB
Editing Systems
Publishing Systems
Rights Management
Order Processing
Archive/Backup
Storage
Hardware/VM
Archive Apps
Hardware
Admin
Hardware/VM
OS
Archive Apps
IaaS and On-Premises Hybrid
Figure 4
11. cognizant 20-20 insights 11
SaaS and On-Premises Hybrid
This hybrid model (shown in Figure 5) combines
the features of a SaaS with third-party services
hosted on premises. As far as SaaS models go,
this is more common than a pure SaaS scenario,
especially for customers that require a degree of
integration with their third-party and single sign-
on (SSO) services.
The SaaS/on-premises hybrid has all the advan-
tages and disadvantages of the SaaS model
previously mentioned. Bear in mind that the
degree of integration with third-party systems in
a SaaS/on-premises hybrid model may be limited.
Some DAM SaaS vendor APIs only provide a lim-
ited subset of the features when compared with
their traditional version. This should be validated
with the SaaS vendor.
Another key point about this model is that all
generic ancillary services (DNS, SMTP, monitor-
ing tools, etc.) supporting the DAM system are
managed by the SaaS provider, and only custom-
er-specific services such as directory services,
enterprise service bus, single sign-on, editing and
publishing systems, rights management, order
processing, etc. are managed by the customer.
Communication between the DAM SaaS ser-
vices and the on-premises systems would occur
via Web service APIs — provided that the SaaS
vendor supports them. Your organization may
also notice when comparing SaaS models that
customer-managed data backup/archive and
direct network connection services are not as
commonly offered by SaaS vendors.
Summary of Advantages and
Disadvantages
Figure 6 (next page) summarizes the advantag-
es and disadvantages of different DAM hosting
models. These values will vary depending upon
the capabilities of the different providers, but on
the whole this table reflects the more common
values seen. Situational values could be either
advantages or disadvantages depending upon
the specific capabilities of the on-premises and
SaaS environments being evaluated.
SaaS and On-Premises Hybrid
DAM System
Ancillary Services
Ancillary Services
On-Premises Data Center
Customer
Managed
Network
Hardware
Admin
Hardware
Admin
Hardware
Admin
DNS
SMTP
Agencies/Distributors Company Site/Workgroups
INTERNET
DAM USERS
PRIVATE CONNECTION
Monitoring
Mgmt. Tools
Directory Services
Third-Party Services
SOA/ESB
Editing Systems
Publishing Systems
Rights Management
Order Processing
SaaS Vendor
Managed
Archive/Backup
Hardware/VM
OS
DAM App Mgmt.
DAM Admin
Storage
Hardware
AdminHardware/VM
OS
Archive Apps
Figure 5
12. cognizant 20-20 insights 12
DAM Service Model Comparisons
DAM Hosting Model Attributes On-Premises
IaaS/
On-Premises
Hybrid
SaaS/
On-Premises
Hybrid
Cost Related
Capital Costs
Expense Costs
IT Administrative Costs
Pay-As-You-Go
Time-to-Market/Deployment
Administration Overhead of Application Stack
Management Quality
Control over the Environment
Infrastructure Competency
Security
Archive/Backup Capabilities
Tightly Coupled DAM Applications
Service Level Agreements
Software Upgrades and Maintenance
System Functionality
Degree of Customization
3rd Party Systems Integration
Network Performance
Scalability/Flexibility
Degree of Vendor Lock-In
= Advantage, benefit over other options = Disadvantage, lacking over other options
= Situational, depends upon environment
A Publishing Organization Use Case
Consider the following scenario: A publishing
organization would like to implement a new DAM
to centrally manage and better organize its dig-
ital assets. The business currently has several
different divisions, each using a different system
to manage assets. The overall plan is to implement
a new centralized DAM system and to gradually
migrate users and assets to the new system.
During evaluations of various DAM products,
the organization needed to understand how to
compare and evaluate on-premises and cloud-
based DAM hosting models.
For this example case, we highlight the following
key issues:
• Cost is a significant factor in the company’s
decision-making.
• The organization has a relatively small IT team
that is fully utilized.
• There is insufficient hardware and/or virtual
servers to support a new DAM.
• The new DAM will require approximately 10TB
of working storage.
• The IT team does not have sufficient working,
archive or backup storage available for a new
DAM.
Figure 6
13. cognizant 20-20 insights 13
Comparison of DAM Hosting Options
• There is a minor amount of network infrastruc-
ture that can be leveraged to support the new
DAM.
• The IT team has loosely defined SLAs, but they
are not monitored, reported upon or managed.
• IT uses a hosted co-location data center and
would implement VMs in lieu of physical
hardware.
• Security requirements are moderate.
• There is a minor degree of custom integra-
tion desired with SSO and some other internal
systems.
• There are 3 million digital assets, mostly
graphics, text, artwork and pre-press content,
with limited video.
• There are approximately 1,000 users, half of
whom are at external agencies and distributors.
• The organization would like to have the first
business unit migrated in the next six months.
Scoring Comparison of DAM Service Models
Figure 7 depicts a model that facilitates the com-
parison of the different DAM hosting options.
It is recommended that you change the on-
premises and weight values (highlighted in blue)
as appropriate for your given environment. All
other values should remain unchanged, unless
the evaluator believes a change is in order. In
our scoring comparison matrix, we have defined
values for the publishing organization use
case. In this particular instance, the compa-
ny holds costs in very high regard, and items
like time-to-market and the ability to select a
hosting location rank low, as reflected in the
weighting values shown. As can be seen, the IaaS/
on-premises hybrid is the preferred solution for
the publishing organization.
On-Premises
IaaS/
On-Premises
Hybrid
SaaS/
On-Premises
Hybrid Weight
Capital Expenditure Avoidance 1 4 5 5
Operating Expense Avoidance 3 2 4 5
IT Staff/Management Cost Avoidance 2 4 4 5
DAM Application Features/Functionality 5 5 3 5
Infrastructure Service Levels 3 4 3 4
DAM Application Service Levels 4 0 4 4
DAM Customization 5 5 2 4
3rd Party Integration Support 5 5 3 4
System Performance 4 4 3 4
Reliability/Redundancy 3 5 3 4
Data Protection Backup/Archiving 3 5 2 4
Security 3 4 3 4
Overall Management Simplicity 2 3 4 4
Scalability/Flexibility 2 5 3 3
Disaster Recovery/Business Continuity Capability 3 4 2 3
Meets Regulatory Requirements 3 4 3 2
Vendor Lock-In Avoidance 5 3 2 2
Overall Control of the Environment 5 4 2 2
Time-to-Market/Deploy 1 4 5 1
Can Select Region Where Data & Systems Reside 4 5 1 1
Scoring: 229 273 223
Weighting: 5 = Very High 4 = High 3 = Medium 2 = Low 1 = Very Low
Figure 7
14. cognizant 20-20 insights 14
DAM Hosting Cost Comparison Example
Cost Comparison Example
Provided in Figure 8 above is a comparison that
shows the differences in costs based on different
DAM hosting models for our publishing orga-
nization example. Please keep in mind this is a
demonstration to show the relative cost structure
differences. It does not represent the actual costs
that a DAM implementation will incur. Also bear in
mind that the services offered among the differ-
ent models are not identical, so a less expensive
option is not always the best choice. Your busi-
ness requirements, your organization’s internal
resource capabilities and your budget will all play
a factor in helping you decide what options best
fit your situation.
This example includes a set of production serv-
ers responsible for various functions of the DAM
system including Web servers, application serv-
ers and database servers, as well as rendering,
search and file transport systems. All systems
in this environment are redundant. Storage
includes NAS for active content, archive storage
for content that does not need to be immediately
accessible and data backup storage. Also includ-
ed is the initial setup labor/professional services
cost, which has been capitalized as part of the
project. Bear in mind, the cost comparison model
does not include any ancillary or third-party serv-
ers or systems.
• On-premises costs: This represents a hosted
co-location scenario where the DAM system is
installed in a third-party hosted data center in
which rack space (including HVAC, power and
physical security) is provided, but the server,
storage, local networking and other related
infrastructure is installed, managed and
maintained by the customer.
• IaaS/on-premises hybrid example costs:
There are no significant cost differences
between a pure IaaS and an IaaS/on-premises
hybrid. As can be seen from these costs, this
is an Op-Ex model with an outlay of capital
that includes the DAM software, and the initial
setup services/labor — which has been capital-
ized in this example as a project cost. Costs
of ancillary and third-party services are not
included in this model.
• SaaS/on-premises hybrid example costs:
There are no significant cost differences
between a pure SaaS and a SaaS/on-premises
hybrid. This is an Op-Ex model. Some Cap-Ex
may be incurred during the initial project
setup or onboarding process, depending
upon the vendor’s pricing structure and the
customer’s needs. Costs of ancillary and
third-party services are not included in this
cost model. Storage and archiving costs are
included in the monthly subscription fee,
but are typically more restrictive in storage
allocation and archiving capabilities.
On—Premises
IaaS/On—Premises
Hybrid
SaaS/On—Premises
Hybrid
Production Servers Sample Environment Capex Monthly Capex Upfront Monthly Upfront Monthly
Physical VM Hosts ,VM, OS & DB Software $123,945 $0 $0 $0 $0 $0 $0
Web Servers (VM) — — $0 $337 $31 $0 $0
Application Servers (VM) — — $0 $1,033 $449 $0 $0
Database Servers (VM) — — $0 $2,066 $134 $0 $0
Load Balancers $15,334 $0 $0 $0 $50 $0 $0
Network devices (routers, switches, firewalls, etc.) $28,400 $0 $0 $0 $0 $0 $0
Shared Disk Storage 20TB (RAW NAS) $94,208 $0 $0 $0 $3,600 $0 $0
Offline/Nearline Archive Storage Capacity 20TB $35,328 $0 $0 $0 $200 $0 $0
DAM Software License/Subscription $50,000 $0 $50,000 $0 $0 $20,000 $5,500
Hardware/Software Maintenance — $2,893 — — $0 $0 $0
Colo Rack Space (with power & 1000Mb/s) $0 $2,500 $0 $0 $0 $0 $0
Data Backup @ $0.026/GB tape)/Disk Snapshots — $52 — — $240 $0 $0
Internet Ingress/Egress Data Fees (GB) $0 — $0 $0 $495 $0 $0
Initial Setup & Config/Professional Services $65,000 — $52,000 $0 — $9,000 $0
Monthly Systems Administration — $3,000 — — $1,200 $0 $0
$412,215 $8,445 $102,000 $3,436 $6,398 $29,000 $5,500
For the purposes of this illustration, on—premises refers to a hosted data center that is managed by a 3rd party.
Above specs are roughly based on a DAM system with 1000 users (depending on usage) and ~3 million pre—press
assets. Workflow processes include ingest, metadata tagging, searching, annotation, approvals, distribution and
archiving. This configuration leverages a VM infrastructure for the on—premises scenario to minimize overall server
costs. This excludes ancillary and 3rd party integrated systems. Please note this is a sample environment and does
not necessarily represent actual costs incurred to implement a DAM solution.
Figure 8
15. cognizant 20-20 insights 15
Looking Forward
The on-premises model tends to be preferred in
situations where a high degree of control and
integration with internal systems is required.
It can also make economic sense in situations
where existing in-house infrastructure and virtu-
alization technology can be leveraged to reduce
the overall capital costs. The on-premises model
is better suited for larger organizations that
have sufficient IT staff to manage the DAM and
related systems. On-premises solutions are not
recommended where capital is limited, or IT staff
resources are tight.
IaaS is preferred in situations where the amount
of capital to invest is somewhat limited and/or
in “greenfield” situations where little or no on-
premises infrastructure exists. IaaS is also ideal
when a high degree of flexibility, scalability, cus-
tomization and control over the DAM applications
is needed. IaaS is recommended for mid-to-large
enterprises that have a modest degree of IT staff
to support and manage the DAM service. While
IaaS has lower IT staff requirements than the
on-premises model, it is not recommended for
organizations that have very limited IT resources.
SaaS is preferred in situations where the organi-
zation has little or no IT support staff and there
is a minimal amount of customized business
requirements of the DAM. SaaS solutions tend
to be better suited for smaller organizations or
workgroups that perform a particular function
within an organization. SaaS solutions are the
fastest and the least expensive way to procure a
DAM service. SaaS is not recommended for large
organizations that require a significant amount of
customization or have a high volume of content/
assets to manage.
The most important piece of advice we can give
is that you do your homework. First off, carefully
evaluate your organization’s business require-
ments, internal technology capabilities and
budget. Secondly, evaluate the DAM vendor’s
products and different hosting options in detail,
and ask lots of questions. With the proper level of
assessment, your organization should be able to
make decisions that best meet its requirements
and budget.
By understanding the pros, cons and resource
costs involved among the different DAM hosting
models, your organization can select a DAM solu-
tion that best suits your needs now, and in the
foreseeable future.
Appendix
DAM Vendor Questions
Below is a list of questions that can be posed to prospective DAM vendors when evaluating their SaaS/
cloud-based service:
• How many customers are using your DAM cloud service?
• How long have you been providing the DAM cloud service as a public offering?
• What are the functional differences between an on-premises installation and your cloud offering?
• Do you support 3rd party integration via Web service APIs or by other means for SSO, directory
services and other 3rd party services?
• How is the cloud solution priced/licensed and how do these costs change as services are scaled up
(more users, storage, processing, bandwidth, etc.)?
• Is the service a multi-tenant environment?
• What components of your cloud service are shared among customers (server, storage, network,
database, etc.)?
• Describe how your technical support process works when using your cloud service.
• How do you provide redundancy of the applications and systems (servers, network, and storage),
physical site, regional distribution, etc.?
• What are your data backup and recovery policies? What is the schedule and type of backups
performed? How are backups tested? Where are backups stored? What is the restore SLA?
• Can we be provided with a backup copy of the repository content if requested? If so, please describe
the options available.
16. cognizant 20-20 insights 16
• Do you have a disaster recovery/business continuity process and if so, what is the recovery time for
our DAM cloud service to be fully functional?
• What are your maintenance windows for performing updates to the applications and systems?
• How is the data secured in your cloud offering, at what layers? (Application, server, network, storage,
physical, access control, etc.)
• Can you provide a description of all customer impacting disruptions of your cloud service in the last
12 months?
• What service levels are provided for Internet availability and bandwidth? Are there restrictions?
• Do you leverage a third party for cloud services (an IaaS provider for example) or do you own and
manage your own data centers? If other parties are managing these services please identify them.
• Do you provide a bulk import service? If so, please describe how this service works.
• Can you provide user logging and tracking activities of the cloud-based DAM application?
• What level of application and system monitoring do you perform on the system?
• What level of application monitoring can you provide to the customer?
• What level of application and system reporting can you provide to the customer?
• Does your cloud solution provide any form of file upload/download acceleration?
• Do you support VPN connections between your cloud environment and the customer?
• Do you support direct connections/private circuit connections between your cloud service and the
customer?
• At what physical locations/regions are your cloud services located?
• Are there options to export the repository and related databases should we need to terminate the
service?
• Is your cloud service certified with any regulatory compliance standards such as SOC, PCI DSS, ISO, etc.?
Glossary of Terms
• Ancillary services: Services that provide core infrastructure service functionality leveraged by the
DAM system, but are not a part of the core DAM system. Examples include directory services, DNS,
SMTP (e-mail) and monitoring tools.
• Co-location: A data center owned and managed by a third party that provides physical space where
a customer may host its computer equipment (servers, storage, networking, etc.). Management of
the computer equipment is typically handled by the customer. Management of the physical facility,
including rack space, power, cooling, network connectivity and physical security, is handled by the
co-location provider.
• DAM administration: This is the administration of the DAM application(s) and includes lightweight
responsibilities such as adding users, assigning permissions, modifying workflows, etc. This could be
performed by an application administrator or a power user.
• DAM applications management: This is the underlying management of the DAM application, usually
performed by IT staff, and consisting of initial application setup, configuration, DBA functions, appli-
cation development, scripting, etc.
• Directory service: This service manages a database of users and associated information including
usernames, passwords, security permissions, etc. A common directory service is Microsoft Windows
Active Directory.
• Enterprise service bus: An ESB provides a centralized and standardized registration and commu-
nication hub to permit diverse applications to communicate using a standardized service-oriented
architecture (SOA).