Why should we think beyond server virtualization for operational simplicity?
What is server virtualization?
Virtualization is a process that divides a physical server into multiple unique and isolated virtual servers with the help of software. Typical enterprise use Type 1 hypervisor a software component, which abstracts physical resources such as compute, storage, and network and make them appear as virtual resources to virtual servers.
Why have enterprises worldwide adopted server virtualization?
When server virtualization was in its infancy two decades ago, Enterprises worldwide began to virtualize their server farms for the following reasons.
While server virtualization is prevalent today, what challenges do enterprises face?
Despite the benefits gained by enterprises through server virtualization, the server virtualization era presents other challenges to businesses. In my experience, challenges arise due to one or all of the three factors mentioned below.
How do siloed teams produce cascading technical effects?
A traditional method of managing IT infrastructure is by creating teams of experts in a specific technology domain, but if these teams don't interplay with each other they lead to poor service to the business.
As an architect, how often do you feel constrained by the decisions made by other technical architects?
Here is a simple example, A network architect prefers a Layer 3 design (no overlay) with ToR topology for a highly scalable network.?
Architects and Engineers make decisions in every silo that impacts other silos. To achieve some success with these kinds of deployments, it is crucial to work in a collaborative manner. How many silos do you have in your organization and how well do they work together?
Do processes that complement siloed teams reduce speed and agility for business?
When IT Infrastructure operations are siloed, a complex overarching holistic process is evident.
领英推荐
Let's explore how a process driven solely by traditional practices can negatively impact consumer experience when handled by separate teams operating in silos.
A manufacturing company that offers shared IT services to its internal business units has a specific process for application owners to request segmentation. Below you see a very high-level overview.
At this company, teams are divided by technology domains, and processes are created to fulfill application requirements. Typically, it takes about seven business days to process a network segmentation request for an application. However, if the application owner fails to provide basic information for the request, it can easily take over two or three weeks.
Do products, features, and strategies need risk and impact assessments?
Siloed operations and processes have traditionally been a consequence of limitations in specific technologies. While products and features enable businesses to achieve a set of requirements there are also inherent risks. It is important to consider them during architecture design and operations.
We can investigate how a management decision regarding a particular feature, made without considering the workload characteristics, has had a negative impact on the delivered service.
A financial institution made the decision to cut costs by using a feature called "Compression" in their storage system. The storage vendor assured the bank that there would be little to no impact on the workloads running on the infrastructure. However, after implementing the change, the bank's critical workloads that are IO-sensitive experienced significant performance issues for several hours. The only solution to restore the storage performance was to create a new storage volume without compression and migrate the affected workloads from the compressed volume.
In the scenario described, there is a risk associated with enabling compression, which is a potential loss of performance. However, this risk could have been minimized if the bank had worked with the storage vendor to analyze the IO profile of their critical workloads and conducted appropriate tests on the compression-enabled volume prior to migrating the actual workload. Maybe a “5W1H” approach prior to implementation would have helped them.
While the above is one example, think about other similar topics mentioned below.
How is your organization addressing these types of issues? Don’t you think a solution that breaks the traditional IT Infrastructure silos and enables a single unified process for businesses to consume Infrastructure services will address these problems?
Disclaimer
The contents of this article have not been authorized or approved by my employer. All opinions are my own.