At Cloud Slam ’11, I hosted a session with ‘clouderati’ luminary Jay Fry (my colleague at CA Technologies and @JayFry3 on Twitter) on “The Hitchhiker’s Guide to the Cloud: Tips for Navigating the Evolutionary and Revolutionary Paths to Cloud.â€
Between bad puns and stretched analogies drawn from the classic Douglas Adams ‘The Hitchhiker’s Guide to the Galaxy’ (aka ‘H2G2’) 6-book ‘trilogy’, Jay and I presented two alternative (albeit complementary) routes to deploying private cloud.
I presented an evolutionary path to cloud computing – building step-by-step on your existing infrastructure and investments; Jay presented a revolutionary path – delivering cloud fast without transforming your existing environment.
You can review Jay’s blog for his side of this fun session; I will run through the evolutionary path here.
In an ideal world we might start over again …
In H2G2, hyper-intelligent pan-dimensional beings (i.e. ‘mice’) had a significant IT investment in the most powerful computer in the universe (i.e. ‘earth’). Unfortunately, when the Vogon Constructor Fleet destroyed earth to make way for a new hyperspace bypass, they also destroyed this IT investment.
The average enterprise may wish to blow up their physical computers too, and rebuild in the cloud. However, most also have a significant investment in IT – including hardware, software, facilities, data, people, skills, and processes. These investments may not be perfect, but they are mission-critical, delivering business value, and prohibitive to destroy.
Just as the destroying earth literally just minutes before it would produce the ultimate question to the answer to life, the universe, and everything made a very bad day for Arthur Dent, so too destroying valuable IT investments to move wholesale to the cloud will make for a very bad day for most CIOs.
These CIOs may also have their own ‘Vogon bureaucracy’ to hold them back. Just as this bureaucracy is unavoidable in H2G2, so too may be CISOs, auditors, and other bureaucrats who are holding back cloud plans with concerns over security, compliance, policy, and politics.
This is where the evolutionary path to cloud comes in.
Step-by-step evolution from virtualization to cloud
The evolutionary path to cloud computing requires a lifecycle approach to the evolution of your virtualization maturity:
- First, consolidation of servers and data centers using server virtualization technologies
- Second, optimization of virtual systems to deal with inevitable problems like VM sprawl
- Third, automation and orchestration to enable scalable services at computer speed
- Fourth, dynamic IT delivers agility, and the basis of a full-blown private cloud
This lifecycle drives maturity from platform-centric infrastructure management, through heterogeneous systems management, to a business-focused service management, while driving business value from initial CapEx reduction of consolidation, through OpEx benefits of optimization and efficiency, agility improvements of ‘hands-free’ automation, and finally delivering revenue gains through dynamic IT.
This evolutionary path allows you to:
- Protect & leverage your legacy investment – in hardware, software, people, and process
- Maintain essential infrastructure – especially non-standard or multi-platform physical systems
- Avoid unraveling services, code, process – avoiding cost and risk, at least in the short term
- Manage skills, staffing, buy-in, and security issues – even if they are more perceived than real
Evolution in action
H2G2’s Zaphod Beeblebrox may have ‘evolved’ two heads and three arms, but there are less bizarre cases showing the evolutionary path to cloud computing in action. For example:
Cadence Designs
Cadence Designs is a leader in electronic design automation (EDA), with 4,500 staff and 12,000 servers in San Jose & worldwide. Facing a $10m+ hardware refresh, low server utilization, slow and labor-intensive tools, they looked to private cloud to improve service and reduce costs. Using solutions from CA Technologies (including CA Server Automation and CA Service Assurance) Cadence Designs delivered dynamic self-service, accelerating their virtualization rollout, increasing server utilization, deferring a significant hardware upgrade, and reducing their costs – all while ensuring high quality SLA achievement.
Qualcomm
Qualcomm is a developer and supplier of digital wireless communications products and services, with $10.4bn in worldwide revenue from 15,000+ employees in 146 countries. Rapid business success resulted in constant demands for capacity and continued ‘build out’ was not an option. Instead, Qualcomm deployed mobile data center ‘pods’ based on Cisco UCS servers with physical and virtual server automation using CA Server Automation. This solution enabled integrated virtualized network and server platforms to deliver new ‘capacity on demand’, automated to reduce costs, improve throughput, and maintain controls. Altogether, Qualcomm estimates this initiative saved around $25M in hardware and data center resources, and spared more than 400 kilowatts of power demand.
Evolution challenges & successes
Evolution from earth inhabitant to space hitchhiker created many challenges for Arthur Dent – from mundane (no decent tea) to life-threatening (becoming ‘the late Dent Arthur Dent’) – that held him back from enjoying his new experience.
So too the evolutionary path to cloud comes with many potential challenges that can cause ‘virtual stall’ – the inability to inability to move the needle on virtualization deployment and maturity that prevents IT from delivering the best possible virtualization and cloud outcomes.
While Arthur Dent never could find a decent cup of Earl Grey, with the right people, process, and technology you can push past virtual stall, and accrue benefits incrementally at every step, with fast wins and rapid returns on your virtualization investment that build toward a long-term cloud strategy.
For example:
- Consolidation – savings from server reductions, reduced rent and facilities, reduced power and cooling impact, improvements to business continuity, and better backup and recovery
- Optimization – operational savings from reduced management and licensing, staff efficiency and redeployment, improved visibility & control, better availability, and improved service mobility
- Automation – greater flexibility, lower error rates, reduced risk, improved staff mobility, and faster response to service requests at ‘computer speed’, not ‘human speed’
- Dynamic IT – faster business reactions to market forces, rapid on-boarding of new partners and channels, better customer service and retention, and faster new product and services delivery.
Key use cases
Marvin the Paranoid Android may have a brain the size of a planet, but for mere mortals, key services suited to the evolutionary approach include:
- Simple, single-server applications – the ‘low-hanging fruit’ of virtualization and cloud, including departmental applications, collaboration servers (esp. SharePoint)
- Constant predictable workloads – such as e-mail and ERP systems, but any system where utilization is likely to be constant and relatively stable
- Non-standard environments – such as UNIX workloads, or any workload dependent on specific hardware e.g. scanners, 3D printers, data input devices
- Highly controlled or managed environments – where Vogon bureaucrats need (for better or worse) to maintain high security, control, compliance, audit, etc.
Summary
Jay allowed me to have the last word, so I used the opportunity to reinforce the message on the front of the Hitchhiker’s Guide – “DON’T PANIC!†As we had heard from Forrester’s James Staten just days earlier at Cloud Slam, few organizations have deployed private cloud yet; most are just starting their planning; so if you are not far along, there is no need to panic.
Nevertheless, you should not delay. With an infinite improbability drive, you may travel through time, visit The Restaurant At The End Of The Universe, check out your cloud future, and return to start your deployment anytime. In the real world, you cannot just wait and hope – you need to act decisively – unlike Arthur Dent – and sooner rather than later to remain competitive in business.
However, it is also important to understand that cloud is not a single destination in space-time. It is not a choice between evolution and revolution; it is not only possible but often advisable to use both approaches – as many CA Technologies customers have done.
So if you haven’t started, don’t panic, but, with an understanding that your path to the cloud, whether evolutionary, revolutionary, or both is going to be difficult, rewarding, and unique.
‘42’
Which brought Jay and I to slide 42 – the answer to the question of life, the universe, and everything, and the ultimate slide in the presentation. So in with the words of the second-smartest creatures on the massive computer called earth (i.e. ‘dolphins’) – “So long, and thanks for all the fish!â€
If we haven’t scared you off yet and you’re interested in more, you can access our Cloud Slam presentation here: “The Hitchhiker’s Guide to the Cloud: Tips for Navigating the Evolutionary and Revolutionary Paths to Cloud.â€