Qlikview as PaaS, IaaS and SaaS Service
QlikView is an in-memory software platform and so speed of
access to RAM is a key factor in delivering good performance to users.
Similarly, availability of CPU power to perform aggregations on the fly and
produce content for users is also key. Virtualization technologies place some
restrictions of speed of access to RAM compared to physical servers and can
also limit the number of CPUs available to a single machine; this is reflected
in the maximum size that cloud servers can be with regard to RAM and CPU. This
can mean that not all QlikView deployments are suited to being virtualized and
so it is important to size a deployment of QlikView when considering use of
cloud.
With QlikView this would mean all users would view the same
QlikView apps but perhaps the data they see within those applications would be
different.
QLIKVIEW AND IAAS
IaaS is a common cloud model used by QlikView customers;
indeed QlikTech’s own demo.qlikview.com and other services are run from the
IaaS model offered by Amazon EC2. This model allows QlikView customers to
deploy infrastructure suitable for their environment and have full control over
what QlikView applications they host in the environment and how they configure
it.
With QlikView in a PaaS model, a customer could implement
QlikView onto a hosted PaaS vendor’s technology. For this type of
implementation, QlikView requires a Windows platform to run on. One such PaaS
option is Microsoft’s Windows Azure. It is possible to package the QlikView
product along with customer - or partner-built apps and deploy onto the Azure platform
and so enable customers to be free of the infrastructure (i.e. hardware,
network and OS) management.
End user access to QlikView hosted on Cloud:
In this kind of deployment there is no customization
required for QlikView, just a simple configuration; however it does require
that the customer has in place an SSO system that can make use of reverse
proxy.In this type of setup the only device that ever physically touches the
cloud server is the reverse proxy. No users will ever directly touch the cloud
server or even know that it is a cloud server. This mechanism could serve both
internal and external users and is an excellent approach for conforming to what
can be the standard way of integrating security within an organization.
SAML (Security Assertion Markup Language) and federated
security such Active Directory Federation Services can be used security systems
that can make use of secure tokens to sign users into a range of systems as
shown below.
With Active Directory Federation Services (ADFS) it is
possible for users to seamlessly log in to a non-domain cloud server without
being prompted to log in as their internal Windows credentials are used during
the process of logging in. This gives an excellent user experience and ADFS is
often implemented within organizations alongside their regular Active Directory
which means there is no requirement for an additional SSO product or set of
users.
In some cases there is no security system to integrate with
beyond perhaps a known list of user IDs and passwords. In these cases a
straightforward login page can be created in a similar way to the method
previously mentioned to ask for, and validate, a user ID and password before
logging them into QlikView. The user directory could be anything like Active
Directory, LDAP, SSO systems, Databases (SQL, Oracle etc), Web based systems
such as Windows Live, Google accounts or Facebook login.
