Werner Vogel just finished his “Ahead in the Cloud” talk here at the DISC 08 workshop. As always, very entertaining. Amazon is light years ahead of the rest of the industry in thinking about and delivering utility computing to the world.
The talk was not technical given the audience here but here are some highlights/notes (subjective of course):
- Werner introduced himself as “a systems administrator for a small bookshop in Seattle” 🙂
- Animoto story (always great hearing about this success story)
- It appears that Amazon has the capacity to deal with MANY Animoto scenarios at the same time, if a situation arises
- Apparently startups are not the only significant load for Amazon‘s services
- 20 billion objects in S3 and increasing exponentially
- Utility computing in the cloud: moves CAPEX to OPEX
- A very interesting point about utilization… if we make it cheap/easy to acquire resources (e.g. provision a server), engineers will not think twice releasing them when they don’t need them anymore hence improving utilization of the infrastructure.
- “Software as a service”: Hmmm… I think I am starting to dislike this term. Werner used it. They are not delivering software but “functionality”. Salesforce, for example, hosts the functionality for our behalf. It’s “functionality as a service” or just “service”. Anyway… just terminology I guess.
2 responses to ““Ahead in the Cloud” – Werner Vogel’s”
“I think I am starting to dislike this term. Werner used it. They are not delivering software but “functionality””
Agree. Actually the diagram you have on a previous post (http://savas.parastatidis.name/2008/07/14/c5f06dce-b4fd-4ba6-b005-7c3899974e7c.aspx) is much more clear. I’ve been working on SaaS for quite some time now and came up with this
http://blogs.southworks.net/mwoloski/2008/07/10/saas-taxonomy-map/
It needs some more refinement (like separating IaaS from PaaS for example) and introduce other dimensions (like behavioral, technical and business value).
The work that you are doing on the Research Output Repository (btw, excelent work) has its place on SaaS probably on the Datamodel extensions capability. And it could be extrapolated to the Platform dimension by exposing the API to ISVs through the cloud (like force.com), or even by using the repository at the enterprise level.
Thanks Matias,
yes…we’ve been thinking about Research Output Repository as a Service. We probably have long way ahead of us 🙂
More info on that front in 10 days 🙂