CloudCamp London, last week, 13th Nov. In the church off Clerkenwell Green near the old Freeserve office that I'd never really noticed.

Generally, a very interesting evening. Heavy infrastructure focus. Clear that cloud models, their adoption and thinking across both providers and consumers, is still immature and developing.
Market growing quickly. Couple of very smart and (becoming) successful products out there.
Sun had a good strategic understanding of where the market was likely to grow from first (more to Paas, Daas, Iaas than SaaS). Plenty of overlap/confusion/misuse/synergy between cloud, grid, virtualised, dense format, and utility computing. In some realms the terms are interchangeable, but as product offerings they are cleary distinct. Exactly what the cloud is remains unclear to many people. The market is young so it feels very open. This is good and bad. ie Opportunity, but market needs to be ripe. We want to start building to be ready, question is whether all the other players are sufficiently visionary.

---------------------------------
Notes from various lightning talks (5 mins), discussion groups & 121 conversations.

Canonical.com - Simon Wardley (consultant / snr manager)
Importance of standards, transparency & portability.
Highly pertinent comparison with the re-packaging of debt in current financial crisis leading to the collapse of the different levels of products. Message being that if you aren't aware of how your suppliers have built their product from cloud & saas, or your suppliers suppliers, then your service can come crashing down if only one saas block in the stack fails. However if you've anticipated this by designing in standards-based portability then you may be able to mitigate.


Quest - Joe Bagley (CTO)
Cloud, grid, virtualised & dense format (blade etc) computing models' green credentials.
Only 0.3% of input power to data centre may make it to CPU. Even idling, huge amounts of power are lost. C/G/V/D approaches are improving things a little, BUT short version is, few models really green. None doing anything much useful to report at granular level on env/carbon budgets. Not one of the cloud providers has proper published environmental position/policy.


Arjuna / Newcastle Uni - Paul Watson (Professor)
Development of Service Agreement & Policy management structures/standards for governing & controlling hand-offs from enterprises to private or public clouds.
Ie proposing that "contracts" should be developed to support cloud federation (but this is currently an academic exercise).


RBS - Rhys Jones (Tech/Infrastructure Director type)
Clouds are cool. But there are obstacles to corporate cloud adoption.
For corporates, clouds can rpovide benefits (see CIO feedback from Cisco) ... move budget from capex to opex (more predictable), and automatic optimisation.
Also represent a rapid application / rapid infrastructure option - which can allow bypassing of normal controls (governance often associated with biz/tech roadmaps & capex control).
[For us, this is the classic opportunity: an organisation wanting to innovate or develop rapidly outside their normal stack because it creates a pseudo-greenfield]
But, there are obstacles in changing the *culture* of control (both governance, and hands on tech control).
Additional point that many corporates wont move until more convinced that the "service gaps" in cloud offerings are being closed.


Cisco - Phil Dean (snr architect / client support type)
Feedback from CIO's on current state of cloud & Saas models/offerings.
Like:
Consistency (support easier globalisation)
Simplicity (easier to 24/7 support)
Service orientation (promotes agility)
Don't like:
Loss of control
Compromised risk management & security
Migration (& hence transitional hybrid operation)
Technology focused, not business focused
Need:
SLA's, standards, security & BCM at "service aggregation" layer.


Zimory - Philipp Huber (CTO?)
There are still things missing to drive to pervasive cloud computing
- interoperability standards
- too much chaos in vendor mkt
- security issues (solution= pervasive encryption?)
- transparent billing models
- SLA's (jury still out on EC2 move in this direction)
Zimory = interesting server resource brokering model ie online market for spare cpu cycles.


Sun - Wayne Horkan (UK CTO)
Sun predicts a "goldrush" to cloud models, though expect this to be dominated by infrastructure oriented Paas, Iaas or Daas.


Microsoft - Neil Hutson (Snr Director - thought he said for Telco, not sure)
Azure. Duh.


Cloudstart - Duncan Johnson-Watt
Cloudcover product - insurance through the cloud.
Tools & dashboards for monitoring & managing an operations different clouds. Enabling (Gui-driven) failover & failback between diverse clouds & images.


Rightscale - Josh Fraser (Biz dev)
Appears to be some very smart bits. Enables full automation of on-demand virtual servers (configured as your image) where you want them when you want them. Either according to pre-defined demand schedules (timeslots), or more cleverly, according to rules based on your existing infrastructure's performance. So, this allows you to automatically "spin up" new resources (e.g. an extra mysql node) when your infrastructure is under stress (cpu, io, ram). Fox news used this approach to bring online new db servers (36 mysql nodes!) to manage peaks in "voting traffic" from their website during US election.

---------------------------------
Other blogs:

http://gojko.net/2008/11/14/cloudcamp-london-2-private-clouds-and-standardisation/