There are a lot of IT people in traditional enterprises all at different levels of understanding and skillsTo become successful with your cloud strategy (implementation) you must have engagement and direction across the teamA bunch of talented individuals working autonomously will not get you the results you need
If your security approach before cloud was loose, your security approach in cloud will be looseBe diligent, diligent and then diligent. Create a questionnaire for “public” cloud service providers / SaaSetc and ensure you fully understand what they do and how they do it. If they won’t tell you, find a new vendor.Don’t be drawn into a base assumption that “the cloud is not secure”. Ask is it “more or less secure than what we have” – work from there.
Despite the hype and the vendor frenzy, not every business or every workload in every business will “need” cloud.Before you embark on a cloud strategy, be very clear on what you are trying to achieve and how it ultimately benefits your business.Defining and executing a cloud strategy is a commitment from the business and IT (if you want to embrace “rogue”) - don’t take half measures.
Most enterprises are NOT Netflix, in fact, around 90% of our 2000 LoB apps have no notion of horizontal scaling capability.Shared Nothing / Eventual Consistency is something that many “traditional” enterprise developers could not speak expertly about.Many traditional enterprise applications are tied to “behind the firewall” concepts for even the simplest things such as authentication (AD) and are not readily capable of even the simplest new approaches, such as federation.
Virtualization is undoubtedly a cornerstone of all cloud, but many people will fall into the trap of believing that highly-virtualized (density) servers = private cloud. Any private cloud that is deployed without the capability of full automation / orchestration is not a private cloud.There is so much “immature” private cloud technology on the market that you will encounter a whole world of difficulty in initial deployment(s) as it “won’t quite meet our needs” and you’ll be back at the drawing board, pronto.There will be too much resistance (not invented here, don’t understand, can’t support) from your Ops guys in the early stages, be prepared to have to explain, time and time again, why cloud is not their enemy.
Ops and Apps guys like things stable. You will come along with a whole new set of ideas and approaches and they won’t want to hear you.Apps guys spend a huge amount of time and money “certifying” their apps for a specific platform / stack. You will come along and want to change this. They won’t want to hear you.You will show them, end to end, how their apps can be provisioned, self service, with a single click. They will assume you are using snake oil.Ops guys won’t want to embrace what you are bringing as it essentially threatens their very existence and their resistance to change is ultimately going to kill them (they don’t get this – yet).
In 2006, I put together a project plan that took our “private cloud” deployment through to 2008 “for completion”. 6 years later, we are finding new ways to create more efficiencies throughout the stack. This is they way we work. There is no “destination”.Cloud, for us, is a way to delivery IT services in ways like never before. This has a direct benefit to our business. We are only now beginning to realize the benefits of what we’ve built, using the “private cloud” and a combination of external services to help drive innovation.
The arguments around IaaS, PaaS, SaaS, Private, Public, Hybrid, etc are largely futile. Understanding the differences between them is very important to your selection of approach(es) but there is no perfect cloud. Every workload should be assessed and fulfilled according to the “best fit”.Expect to become “service aggregators” – the days of “nobody got fired for buying IBM” are over.Be clear that as you move to a service aggregation model, your support mechanisms and the organization for providing support will need to change – significantly.
- Simple take away – “just because you can do something, doesn’t mean you should”. Cloud, to me certainly, represents a tectonic shift in the way we provide IT services to enhance our business. It’s not solely about virtual machines, software defined networks or infinite scale, they are just components, as components of an engine simply work together to make a car run.
1. CloudCamp North Five Alive Christian Reilly – Bechtel Corporation
2. Disclaimer / Cover My *aaS All the rubbish you’re about to hear is based on experience and opinion.
3. Who are you? @reillyusa on teh twitterz Loose Couple Blog (http://blog.theloosecouple.com) Sort of like a biz > tech guy in a massive “enterprise” company Bechtel is the largest privately owned E&C company, evar Accidentally designing “Clouds” since 2006 mainly in the USA Manchester City FC Season Ticket Holder
4. It’s Not The Individuals, It’s The Team
5. Blah. Blah. Security Generally, it’s the #1 excuse for not moving to cloud, not the #1 reason.
6. Bad securitay is just like bad, m’kay ?
7. Five Alive - #1 Don’t do cloud just because you think everybody else is.
8. Not All Things Are Made Equal
9. Five Alive - #2 The typical enterprise has very few applications that suit public cloud models.
10. Complexity = Perplexity
11. Five Alive - #3 Your initial private cloud effort will fail miserably for a variety of reasons.
12. Your Private Cloud – Version 1
13. Five Alive - #4 Be prepared for long and epic battles with ops guys and application owners.
14. Honey Badger Don’t Care
15. Five Alive - #5 Your cloud deployment will never be complete, not in your career, anyway.
16. The Long And Winding Road
17. Five Alive - #5.5 There is no perfect cloud. Use the best fit for each of your target workloads.
18. Often Asked Question “What is a cloud?”
19. Often Given Answer “Does it really matter?”
20. I Think It Looks Like This
21. Final Thought Cloud is a journey. A new approach and a conduit for business transformation.