#ActionItem - CrowdChat

AWS re:Invent 2018 Highlights

What happened at AWS re:Invent 2018? Hear highlights from theCUBE's 100+ sessions. Then, join us for Wikibon's AWS re:Invent 2018 community review.

#actionitem @wikibon @plburris @stu @sarbjeetjohal @jameskobielus @dvellante @dfloyer

LeaderBoard

#actionitemAWS re:Invent 2018 PreviewAWS sets the pace in cloud computing. AWS re:Invent is best place to learn AWS's speed and position. Join us for Wikibon's AWS re:Invent 2018 community preview.

#actionitemHybrid ApplicationsHybrid applications promise a powerful way to distribute digital services to improve customer engagement and business operations. But how will they work?

Stream Ended

Establishing a secure connection

The cloud is reshaping technology choices, development paradigms, and tech business models. How do you see it affecting vendor/IT engagement practices? https://www.crowdchat.net/s/85qag

[+] Show Hidden Comments

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

Engagements are going change to how does this choice affect what I deploy in the cloud today - and how I manage my cloud. This will traverse Management, Devops, security, storage and data. If they do not accommodate how I use the cloud - then they will not be a viable option

Cloud is making self-service the IT/vendor engagement model for even the larger, more complex buys. Going forward, it's hard not to see AWS, Azure, and others scaling back direct sales as buyers using self-service marketplaces for bigger-ticket items.

traditional vendors have no choice but to pick cloud partners - they bring a lot to the table in terms of customer relationships and services but the historic intel and msft partnerships are morphing to cloud partnerships.

The gap between when stuff get produced/released to when it is consumed is shrunk to almost zero days. Old days of releasing things in half-hazard way are a thing of the past. Your service is your press-release ;) Old guard need to understand this clearly!

The vendor customer relationship will be automated away and value relationship will ensue after - a new model will emerge imho

Any vendors which don'y play well with mega-platforms (AWS Google Microsoft Alibaba Tencent) will suffer.

Old guard vendors must invest heavily into marketing for developers... A must ....

What re:Invent 2018 announcement -- from AWS or anyone else -- was especially “ho-hum?” https://www.crowdchat.net/s/65q7m

[+] Show Hidden Comments

I think there were plenty of bigger, better faster announcements that were incremental - but unlike most vendors it was a small proportion of the corpus of announcements at #reinvent2018

- “Well Architected” effort, so far seems like a marketing stunt, or it might be that I don't understand how effective that can be... or how that part can be automated... is there an API to compare my architecture with a "good architect"

AWS Outposts. "Ho-hum" in several senses: 1) it's not going GA till late 2019, 2) it comes late into the market vis-a-vis Azure Stacks, 3) AWS already has hybrid cloud connectors & running VMware on AWS. I'll get excited about Outposts if it, per David F., is huge hit.

This is the kind of question ussually reserved for events with a few announcements but AWS has a power law of announcements from big to the long tail all relevant imho

it's hard to come up with an answer to this one ;)

I think that the if we had to be critical I think the blockchain was one that was mainly vapor or more market pleasing check box

Undoubtable the Edge. AWS have realized that 90%+ of the compute will be dedicated to inference code from AI models. Real-time inference code is the most valuable, and will need to be moved next to where the data is created, almost always at the Edge.

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

I would say deepracer - it was cool - but what is the business use case for such a service / device ?

I also wasn't terribly impressed by AWS Ground Station. It's not clear to me whether they and Lockheed Martin will be able to market that effectively to the SMB market that perhaps could benefit most from access to remote sensing data etc.

@maishsk It's a prototyping platform to prove out the RL features of SageMaker RL and the simulation capabilties of RoboMaker.

Abner @reinvent

@jameskobielus I don't think Outposts needs to be HUGE as in lots of people use it. It's huge if it accelerates AWS's top 100 global customers to go all in on AWS. There will also be a mis-match between AWS regions and global needs. Outpost solves that and 10 other things.

Kenneth Hui @rubrikInc HQ

The DeepRacer made a big splash initially. But in talking with fellow attendees, most of them thought of it as a free toy for themselves or for their kids. :)

@abnerg I don't think "going all-in" on AWS is a business requirement that Outpost "solves."

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

@jameskobielus So it is a nice toy that can solve an AI problem - I understand the logic behind it - but announcing a new "product" just to prove the point - to me seemed flaunting their dominance.

Captain Dee 🧜‍♂️

@abnerg if AWS achieves 10% market share they will be happy. If they hit 50% it will completely disrupt the market. At this point it's just value add for them. Deployment options.

Given your assessment of the cloud market “as is” and your expectation for the cloud market “to be,” what technology segments are most underserved by cloud today? https://www.crowdchat.net/s/25q81

[+] Show Hidden Comments

- Network enablement for Hybrid Cloud still is an underserved area. Closely related is data movement between cloud and private data centers. Gluing of pipes and widening of pipes is still needed.

Undoubtable the Edge. AWS have realized that 90%+ of the compute will be dedicated to inference code from AI models. Real-time inference code is the most valuable, and will need to be moved next to where the data is created, almost always at the Edge.

- NFV (network function virtualization) is another weak area (in #HyBrid context)

Intercloud networking. VMware NSX is a guidepost for how cloud networks will get "flattened" over the next few years.

- Policy based computing is good in compute stack, weak on storage and weakest on network side. Network is most closely related with #security also, that’s where more investment dollars need to go. Policy based network security. #Cisco?

Captain Dee 🧜‍♂️

all of the boring transnational processing done on old IBM servers in basements.

Captain Dee 🧜‍♂️

@dfloyer 5G and advancements in internet and networking will negate the need for the Edge. The two best Edge case studies I've seen included deployment on a container ship and cruise ship.

In terms of what solution segments are underserved by public cloud offerings, I'd point back to business intelligence. Yes, AWS, Microsoft, Google, IBM, et all have SaaS-based BI offerings, but somehow they still haven't penetrated the mass of business users.

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

I would say - edge computing - AWS is only starting to scratch the surface of what can be accomplished today

Serverless has lot of growth as this combined with containers makes lots of stuff happen

Assume that cloud invention and adoption sustains its pace, what will be the role of an IT organization in 2023? https://www.crowdchat.net/s/55q98

[+] Show Hidden Comments

Kenneth Hui @rubrikInc HQ

Get out of the way? ;-)

Captain Dee 🧜‍♂️

Devops

@kenhuiny be a tailwind...no blockers...facilitate innovation...be an expert in applying services to solve business problems

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

To stay relevant - IT have to be an enabler to the rest of the company to move to the cloud. Absolute minimal intrusion into speed of development - but still guarding the company's "crown jewels"

In 2023, the role of an IT organization will be to assess the optimal allocation of enterprise assets (data, workloads, apps) among clouds and on-prem, to manage/secure it all, and to ensure consistent service level across multicloud.

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

To stay relevant - IT have to be an enabler to the rest of the company to move to the cloud. Absolute minimal intrusion into speed of development - but still guarding the company's "crown jewels"

Kenneth Hui @rubrikInc HQ

Seriously, IT orgs need to take a service provider approach where they create sensible policies for users and focus more on services availability instead of component availability.

The role will fundamentally remain the same. The tools will be different. Time spent on infrastructure will dwindle, time on Digital Business dominant.

IT becomes important as a governance body! Keep the LOBs honest... #Ethics of technology (especially #AI)

People + Machines do better than manual or automation alone. #2MA #reInvent2018

As enterprises make more formal and strategic moves to the cloud in 2019/20, what considerations will dominate cloud vendor selection? https://www.crowdchat.net/s/65q8q

[+] Show Hidden Comments

Cloud-first architectures for hybrid cloud will dominate, and will be the basis of what Wikibon calls “True Hybrid Cloud”. The key capability is the ability to dynamically move code without recompiling or testing.

Kenneth Hui @rubrikInc HQ

The cloud provider who can make the transition the easiest. That includes pricing that fits enterprise procurement models and prescriptive approaches that give enterprises options but not an infinite amount of options.

- #1 Criteria is Price/Performance ratio and it's is more important than vendor lock in. #2 is functional depth of cloud service provider (innovation velocity). #3 is vendor lock in avoidance. #4 is usage of existing skill-sets from existing workforce #5 value

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

The acceptance that hybrid is here to stay - and the vendor that can accommodate that journey in the easiest - until everything moves to the cloud - will win in the long run.

The primary consideration for enterprise selection of cloud vendors in 2019/2020 will be the range of hybrid-cloud options provided: direct connection/interop between public/on-premise + put private assets in public + put public cloud services on-prem.

Captain Dee 🧜‍♂️

simple: ease of use, costs, and platform/development options.

Providers which does hand holding in applications portfolio analysis and helps in migrations will do great.

Data movement among clouds. Ingress shouldn't be 10%-20% the cost of egress, especially with hybrid, on-premises offerings.

Within the context of flexible hybrid-cloud on-ramps, whether a cloud provider offers a seamless cross-hybrid experience, including fast/cheap data/workload migration, will be another criterion.

A provider which shows Legacy Workload Appreciation (empathy) will do better as well!

Other people with Cloud-first hybrid architectures include Azure Azure, and Oracle Cloud@Customer. AWS has legitimized this approach.

Wikibon will be pushing out an aggressive True Hybrid Cloud forecast in the near future.

Captain Dee 🧜‍♂️

@sarbjeetjohal agree 100%!

What will be the first market impact of AWS’s Outpost hybrid cloud offering? https://www.crowdchat.net/s/65q9h

[+] Show Hidden Comments

Product won't ship for 1 year so we'll see what signals will emerge. My guess is we'll start seeing what RDS does with VMware as a proxy

The first market impact of AWS Outposts, when it ships in late 2019-ish, will be on the AWS' public cloud rivals. Everybody will hook up with hyperconverged infrastructure vendors to deploy equivalent hybrid-cloud appliances.

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

If it does what I think they are planning for it - we will all look back at the announcement at #reinvent 2018 - as the day that AWS completely dominated market - in the datacenter - and in the cloud.

It will legitimize the market - Cloud-first architectures with identical hardware and software will dominate hybrid cloud and an important component of multi-cloud growth.

Proximity Wins! Latency Matters a lot more than what we talk about! First and foremost, it legitimizes the need for private cloud to cloud bridge (cloud in a box). It's a good as well as bad news for incumbent vendors!

first market impact of Outposts is all the competitors will say it validates their on-prem strategies - but behind the scene it's an "Oh shi*" moment that will freeze the market imo #reinvent2018

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

@dvellante Why do you think it will freeze the market?

Kenneth Hui @rubrikInc HQ

I expect Google Cloud will respond with their own "cloud-in-a-box" solution. They announced an on-prem solution at Next this year but it was for Kubernetes and not for other GCP services.

Kenneth Hui @rubrikInc HQ

Or a Nutanix/GCP solution will be what Google pushes as their answer to Outpost.

Maish Saidel-Keesing (☁️🚀☁️🚀☁️)

@kenhuiny Aren't they already doing that with Cisco https://cloud.google.com/blog/topics/partners/cisc...

Kenneth Hui @rubrikInc HQ

@maishsk Still seems Kubernetes specific an not GCP services on-prem per se.

@maishsk I feel that way because every VMware customer considering an on-prem/hybrid solution now has another choice that won't ship for a year.

@maishsk so I think it adds confusion to the market and forces tough questions