AWSreinvent

AWSre:Invent Influencer Notes
A VIP Influencer Chat. Unofficial off-Hashtag chat for influencers. Show's official tag #reinvent
   8 years ago
#awsreinventAWS re:Invent 2015 PreviewThe latest from the cloud crowd
John Furrier
AWS cloud competitors are all touting and claiming massive cloud growth. Is it really cloud growth or just other legacy products bundled with cloud? How do customers know what "real" cloud is?
ConsumingML
depends on your definition of cloud. #AWS tells you massive scale w/ elasticity is cloud. CIO, VP, & Company Directors tell you its just the ability to not run on-prem.
Muddu Sudhakar
Legacy apps and infrastructure spend is going towards cloud vendors; which is driving growth for all of them;
Muddu Sudhakar
The revenue should be defined as Cloud+SaaS; for example, Microsoft is claiming SaaS apps (Office 365, Exchange 365 etc) revenue in Cloud business; And Google is including SaaS apps/docs revenue in GCP; \
John Furrier
there is multiple clouds developing #SaaS clouds like from ISVs and companies like Oracle so this is an interesting question to flesh out. Is there a thing as a "single cloud"
John Furrier
Where do you see the value that cloud creates for customers in the next 24 months #cloud ?
Subbu Allamaraju
agility from the service ecosystem, and rapid second-order architecture changes (thereby clearing tech debt)
ConsumingML
@sallamar @Furrier agree on clearing tech debt. Born on Cloud apps are already here. Next wave, migration of existing apps.
Muddu Sudhakar
Legacy SaaS applications will be disrupted by Cloud, Micro-services, Serverless and AI/ML
John Furrier
How will #AWS take its early lead in serverless (with Lambda) into broader usage? There are a growing number of open source options for serverless which would be attractive to developers which are the most relevant?
Subbu Allamaraju
Open source options don't count as much due to the strong role of service ecosystem. See (3) on https://m.subbu.org/... (don't mean to plug)
yaron haviv
most FaaS r docker/k8s add-ons, not a self-service platform (logging, security, data,..) address dev not ops
John Furrier
@sallamar great article on serverless!!!
Keith Townsend - Light will overcome darkness
I did a #CTODailyDose on this topic. to @yaronhaviv point. Open Source solutions address FaaS but not Serverless. Big debate in the CNCF Serverless working group as to what's Serverless.
Keith Townsend - Light will overcome darkness
Leaves opportunity for multi-cloud solutions. I don't believe AWS has won Serverless. Far from it. CF isn't FaaS but it is serverless when operated by someone else.
John Furrier
The CDO experiment starts to wind down. CIOs take over in charge of data and start to focus on data value, which bolsters their standing -- again. True or False? Comment. Is the CDO fade over?
John Furrier
The industry struggles to distribute authority in systems. Concerns about systems of agency actually emerge as an issue in the 2018 election. True or False? Comment
John Furrier
Enterprises start taking specific steps to reduce their investment in legacy maintenance, which today is 35%+ of spending. By 2027, today’s legacy will capture less than 15% of total spending. True or False? Comment
ConsumingML
False for S/W. Legacy will just morph from on-prem code to stringing services together code. % doesn't change, just will be able to do more with the 35%.
John Furrier
Some say False bc many legacy systems and solutions will be updating with AI and ML so that might keep the spend high
John Furrier
#ML based ITOM will grow exponentially. It’s one of the first horizontal packaged apps for ML. And customers will recognize it as “training wheels” for future IoT applications. True or False? Comment
John Furrier
IoT will go sideways in 2018: Too expensive, OT & IT issues, waiting for volume solutions from new vendors, and IoT volume vendors will use low-cost infrastructure. True or False. Comment
John Furrier
No consensus will be reached on logical location of IoT edge analytics or persistence of data. True of False. Comment
John Furrier
90 per cent of machine-learning algorithms and their code are now available open source through various communities (e.g. TensorFlow, Scikit Learn, H2O and Weka). True of False. Comment
ConsumingML
True - higher actually w/ exceptions around specific financial apps. All general purpose algos in OS. Unless trading on an exchange, you are using Open Source.
Subbu Allamaraju
Either open source, or commoditized via services on public clouds
yaron haviv
right, but the challenge is getting the learned models and acting on it