hybrid cloud infrastructures: 9 circles of hell to avoid
DESCRIPTION
There are prominent challenges that enterprise end-user organizations face when building private clouds and the common pitfalls of attempting a private cloud project. Xangati listed out the top nine common pitfalls to watch out for so as to avoid a failed private cloud implementation.Know more about Xangati: http://xangati.com/TRANSCRIPT
-
Hybrid Cloud Infrastructures: 9 Circles of Hell to
Avoid
Xangati Blog
Atchison Frazer Vice President, Marketing
June 17, 2015
-
Xangati Blog
The Gartner IT Operations Strategy & Solutions Summit wrapped up here in Orlando today,
and having attended this event many times in the past, it was striking that the title changed to
include strategy and that the majority of survey responding attendees cited strategy as a key motivator for attending.
One of the strategy sessions that I attended was, Why Private Clouds Keep Failing, presented by Gartner research director Mark Lockwood, who previously was with Eli Lilly and
Company, where he served as virtualization architect and end-user computing architect. Mark
described the top challenges enterprise end-user organizations face when building private
clouds and the common pitfalls of attempting a private cloud project. He analyzed the
shortcomings of common cloud management market offerings and offered recommended
steps to mitigate those challenges. He used a bulls-eye graphic playing off Dantes nine circles of hell/fail), and it seemed to me, especially given Marks background, that there is a strategic role for infrastructure performance management tools to play to beta-test private/hybrid cloud environments. So, with apologies to Mark (and Dante), here are my nine
circles of hell to AVOID a failed private cloud implementation.
-
Xangati Blog
1. Policies Team/Structure: Replatforming legacy infrastructures to a private cloud
architecture naturally threatens to break down traditional silos in terms of roles and
responsibilities, as well as traditional infrastructure topologies. Virtualization infrastructure
admins are most likely the closest profile within IT to matriculate to private cloud design
(humbly stated by a longtime networking guy). What better way to coalesce the disparate
IT functions that a private cloud architecture overlays than to standardize on a global view
of all key performance data and efficiency metrics that can help evaluate a business case
for the project? That is especially true if its a top-down decision with procurement dollars already allocated.
2. Process/Governance: There are a lot of pilots in the cockpit of private cloud
deployments. With shadow or shallow IT activities rampant, how do you determine if a
new functional component is out of process or threatens to break the private cloud
architecture? Having an extensible performance monitoring platform, with a shared
risk/reward view of all existing and net-new components that are dropped inline, is a
valuable service to ensure proper governance and to ensure that SLAs are not violated.
Additionally, SaaS apps and BYOD entities that threaten overall process control can be
tracked and reported at a highly granular level to determine of any one of the virtual
service consumption metrics is out of policy.
-
Xangati Blog
3. Automation Complexity: Self-service, rapid provisioning and dynamic workload
scheduling are all desirable elements of a private cloud. But if performance should break
down for whatever reason, shouldnt that be automated too? And I dont mean filling out a service ticket form; nobody wants to have to do that. How about visual trouble tickets that
paint a picture of performance degradation at the exact moment its happening with historical trend analysis of all conventional silo interactions?
4. Third-party Integration: Third party solutions, such as cloud management platforms, will
be required to plug in to your private cloud, and not uncommonly, more than one due to
the complexity of implementation, integration, customization and maintenance.
Performance management can help reduce the complexity of how you gain insights into
systems management, virtualization platforms and cloud management; a 360-degree
holistic view platform, that can take API calls from other management tools, is key to
streamlining third party integration.
-
Xangati Blog
5. Long-term Commitment: While demanding long-term SLA commitments, its important to work with vendors who can also align with your enterprise business objectives. Often it
takes months and sometimes years to fully realize the ROI benefits of a private cloud
implementation, and theres no one turnkey solution. However, for the short term, its important to adopt a performance management solution that offers immediate time-to-
value, that gets smarter the more it learns from the environment, and that can dynamically
flex to the elastic promise of cloud service provisioning.
6. Public Cloud Compatibility: Agility is certainly a goal of private cloud adoption,
particularly with demand for faster stand up of new workloads or services delivered via the
cloud. However, as capacity demand or new market opportunities dictate, some of those
workloads or services might make better sense to run in a public cloud, where cost is
perhaps lower and more easily tracked for chargebacks or show backs. However,
private/public cloud services can often result in cloud sprawl if not properly managed in terms of performance enhancements and overall operational efficiency of on-premise
private cloud infrastructures.
-
Xangati Blog
7. Team Skills: The one IT professional profile that you can demand top dollar for today is
data scientist. One reason is that with the explosion of big data and things that ping other things over the Internet, a scientific approach to data is certainly required. However,
how about traditional IT personnel who want to expand their competencies into the data
realm without securing a PhD? Performance management tools make that possible by
automating the way data are curated and correlated, with dynamic triggers for storm
contention severity and prescriptive actions for more intelligent operations and right size
capacity planning.
8. Infrastructure (Hypervisors, Servers, Storage and Networking): One area to avoid in
terms of Day 2 operation of the private cloud, is to have to log into several different
management and performance consoles for each of the conventional infrastructure silos
that underpin private clouds. If you have a performance management platform that
captures highly granular data not only from the functional components of the
infrastructure, but that also provides a real-time predictive analysis capability across all of
the silos, then your time and productivity investments are better protected.
-
Xangati Blog
9. Service Assurance: Service assurance, guaranteeing end-user quality of experience of
private cloud services including application response metrics, is the ultimate responsibility
for the IT department and a major resource consumption concern of business
stakeholders ensuring that all applications are healthy, workloads are balanced to achieve peak efficiency in the use of the IT infrastructure, and that service-level targets
are consistently met while end users enjoy a superior workspace experience from any
client or mobile device anywhere in the world.
Visit our Blog for more information