1. Avoiding Cloud Computing
Planning & Implementation
Failure
Vancouver Enterprise Cloud Computing Users Group
Meet-up: December 9, 2014
Facilitator: Nathaniel Payne, Senior Data Architect, Cardinal Path
npayne@cardinalpath.com, njpayne@sfu.ca, npayne9@mail.gatech.edu
Image: http://www.google.com/about/datacenters/gallery
2.
3. What Is Cloud Computing?
• “Cloud computing is based on utility and consumption of computing
resources. Cloud computing involves deploying groups of remote
servers and software networks that allow centralized data storage and
online access to computer services or resources. Clouds can be
classified as public, private or hybrid.”
(http://en.wikipedia.org/wiki/Cloud_computing)
Image: http://www.google.com/about/datacenters/gallery
4. Cloud Computing Service Models
• SaaS (Software As A Service)
• CRM, e-mail, virtual desktop,
communication, games, CRM, ERP,
Collaborative tools
• Consume It!
• PaaS (Platform As A Service)
• Execution runtime, database, webserver,
development tools, decision support,
streaming applications, web
• Build On It!
• IaaS (Infrastructure As A Service)
• Virtual machines, servers, storage, load
balances, network, caching, legacy,
networking, security, system management
• Migrate To It!
Image: http://www.google.com/about/datacenters/gallery
5. Is Cloud Computing A Paradigm Shift?
• For most enterprises cloud computing requires a massive shift in
technology, resources, and mindset
• One of the reasons is because, with cloud computing, different lines
of business can now look at building their own technology solutions
without always being constrained by IT capacity.
• Thus, the problem because demand driven & brings new challenges,
including the need for governance, vision, and technology leadership.
Image: http://www.google.com/about/datacenters/gallery
6. What Challenges Do Enterprises Face?
• Inability to fully understanding the core requirements of a problem & trouble translating those
requirements into decision regarding the optimal cloud solutions
• Ask yourself: Do you understand what problem needs to be solved? Who are the key stakeholders?
• Lack of understanding of cloud computing technology & uncertainty with the steps required to
properly transition current resources and technology to the cloud
• Ask yourself: Have you let popularity influence or determine your cloud solution / architecture choice?
• Inability to define a solid business case & map that business case to key post deployment metrics
• Ask yourself: Have we truly considered short term and long term ROI? Have we properly defined the
business cases we are trying to solve so that we can choose the right cloud solution? How will we measure
success?
• Failure to consider the larger organization impact of shifting to the cloud
• Ask yourself: Have we put in place a solid strategic overall vision to help guide our transition? Who will have
responsibility for things in the months after the transition? Years?
7. Advantages Of Working In The Cloud?
• Cloud can bring lots of value!
• Direct and indirect cost savings (servers, buildings, etc.)
• Lower servicing costs? (Ideally… )
• Lower licensing costs for software components? (Ideally… )
• What about unexpected costs? Transition & development costs?
• Often, companies underestimate the effort it takes to build software in the cloud
• Other benefits … Increased agility!
• Business can rapidly expand or contract, and add or alter core business processes, as
required by the business.
Image: http://www.google.com/about/datacenters/gallery
8. Causes Of Failure & Solutions
Image: http://www.google.com/about/datacenters/gallery
9. Cause 1: Wrong Success Metrics
• Inability to create metrics to define the success of the cloud computing
project (Q: How do you define if a cloud computing project is successful?)
• Solution: Focus on the core strategic value points, such as the ability to expand and
change, and less on the “capital expenditure vs operations expenditure” discussion
• Don’t migrate applications to the cloud simply so that you can reduce costs!!
• Remember… Many applications are NOT good candidates to move to the cloud in their
current architecture… That doesn’t mean they can’t be in the cloud... It just means they might
not need to be.
Image: http://www.google.com/about/datacenters/gallery
10. Cause 2: Lack Of Talent
• Inability to locate the right cloud computing talent. (Q: Where do you
find the talent that you need to be successful?)
• Solution: Hire in advance of need. Many enterprises like to leverage just-in-
time hiring. That rarely works in the cloud computing space. Also, be
prepared to pay for the right people.
• Deploying, monitoring, and maintaining software in the cloud can be
drastically different from how an organization currently handles those tasks.
You need people with a broad knowledge of networking, security, distributed
computing, SOA web architectures, and much MUCH more. Develop
internally, and reward expertise handsomely. It is in very high demand!
Image: http://www.google.com/about/datacenters/gallery
11. Cause 3: Weak Governance
• Governance in this context can cover many facets, including strategic decision
making, cost management, security, compliance & planning (roadmaps)
• Inability to properly define and implement the right governance and
security solutions. (Q: What kinds of governance systems support
the cloud deployment and data? Security?)
• Solution: Security and governance are critical & should never be
afterthoughts!
• Never forget to plan for outages & unusual events. Even in the cloud,
everything can fail!!! If this happens, the architecture, and not the cloud
provider, is to blame!
12. Cause 4: Inflated Expectations
• New cloud computing services can sometimes fail to live up to the
expectation of the enterprise! (Q: How do you set and communicate
expectations of the deployment?)
• Solution: You must always test out ANY cloud solution completely before
implementation! All public IaaS and PaaS clouds are very different and have
different strengths and weaknesses. Also, work hard internally to set clear
expectations.
• Also, don’t believe that your organization will suddenly become a digital enterprise
over night. Your organization is a complex one with many vendors and technologies,
as well as likely every software development & architecture pattern that has been
popular over the years. Thus, start small, using the cloud to deliver small projects
that show business value & incremental progress.
13. Cause 5: Misunderstanding Current Architecture
• Many organizations fail to understand their larger enterprise
architecture and fail to understand how their use of cloud computing
technology fits into the larger picture. (Q: How similar is the cloud
architecture to your non-cloud architecture? Different? How do you
manage it today? How would you like to manage it?)
• Solution: Work hard to always understand your current organizational
architecture and sure that you avoid pushing down new and innovative
technology by pointing to the importance of architecture. Also, NEVER
under-estimate the impact that cloud computing will have you’re your
organization!
14. Cause 6: Focus on Technology Rather Than Requirements
• Often, enterprise leaders focus too much on the technology, and not
enough on requirements. (Q: What does the requirements phase
look like for these types of projects? What should it look like?)
• Solution: Never think that moving to the cloud is simply an “AWS project” or
“Google project.” And, to that point, never let popular cloud technology
drive the project requirements. This can be catastrophic, leading to
misunderstood customer requirements. To succeed, often a very strong
technical lead / CTO who understands both the business side of the
equation and the technical cloud side. This person can provide the thought
leadership required to guide these projects to completion.
Image: http://www.google.com/about/datacenters/gallery
15. Cause 7: Good Performance Is Undefined
• In too many cases, performance is not considered until it’s too late.
(Q: How do you decide what good performance looks like for a
deployment? Business case? Does time matter?)
• Solution: Always be sure to define the key success metrics that the project
and its outcomes will be based around. Also, set clear deadlines and push
aggressively forward towards those. How to define success? There are many
metrics to be considered when it comes to the cloud. Some of these,
particularly on the technology side, include:
• Availability, scalability, extensibility, security, architecture match, performance to
specification, latency, throughput, accuracy, cost / ROI, match to developed business
cases
Image: http://www.google.com/about/datacenters/gallery
16. Cause 8: Mistakes During Vendor Selection
• Vendor selection? (Q: How do (or did) you select your cloud
technology vendor? What influenced you? Why?)
• Solution: Don’t select your favorite vendor. Select the RIGHT vendor! Also,
never let loyalty to a current vendor get in the way of the right decision for
your problem! Also, be sure to always know the differences between SaaS,
PaaS, and IaaS, as well as the business cases and applications that each are
well suited for!
17. Cause 9: Poor Security
• Never forget about security! That said, don’t buy into baseless
rhetoric! (Q: How do you deal with cloud security?)
• Solution: Always ensure that your staff understand cloud security (if you are
going to a cloud solution or already on one). If they don’t get a consultant.
Unfortunately, too many enterprise organizations have staff that don’t fully
understand cloud security. In fact, many senior enterprise staff lack an
understanding of the standards or lack experience. Remember, with the
proper security architecture, the cloud can be more secure than most on-
premises data centers. That said, few corporations have individuals who
know enough about current security requirements, protocols, or threats.
Moreover, many are not equipped with the skills to architecture & engineer
for it! (Future Meet-Up Topic)
19. About The Vancouver Enterprise Cloud
Computing Users Group
• This community is dedicated to the architecture, deployment, scaling, activation, and operation of enterprise
cloud computing systems, as well as all the various technologies and applications that these systems can
(and will in future) support. While discussions about cloud technologies abound, enterprise cloud computing
needs its own focused community. This is because issues including architecture, technology, business
impact, internal dependencies, budgeting, security, governance, and application set-up, must each be
uniquely addressed within the enterprise cloud computing environment.
• Without question, this community will begin and remain vendor agnostic. That said, the community will
explore and focus on a variety of Enterprise cloud solutions, including Google's Cloud Computing offering
(Compute Engine), Amazon Web Services, CloudStack, Microsoft Azure, Joyent, SAS, OpenStack, Rackspace,
and others. The group will also be dedicated to developing a best practice knowledge base relating to the
deployment of large scale enterprise applications, including database solutions using Cassandra, Hadoop,
Teradata, MongoDB, Netezza, Vertica, and HANA, analytics and data science applications using SAS, R, SPSS,
or PMML, and visualization solutions using tools such as Tableau.
• Monthly discussions will also review relevant enterprise use cases, and explore the best practice cloud
deployment of large scale analytics systems, data visualization tools, and even ad buying systems. And, while
there will be a persistent technical emphasis, anyone who is passionate about enterprise cloud computing
and large scale application deployment, including business and software leaders, big data developers,
computing researchers, students, or other professionals seeking to learn more, are encouraged to join,
collaborate online, as well as attend future in person sessions!
http://www.meetup.com/Vancouver-Enterprise-Cloud-Computing-Users-Group
20. See You At The Next Meet-up!
Vancouver
Enterprise
Cloud
Computing
Users Group
http://www.meetup.com/Vancouver-Enterprise-Cloud-Computing-Users-Group