HybridCloud

automation

Major monitoring tool gap right now. Monitoring and managing ephermality of instances and data. Most IT tools designed for monolithic and stateful.

hey @dvellante that last developer thread nailed it on most of the reasons https://www.crowdcha...

Cloud Orchestration, Cloud Brokering..

dare I say #kubernetes

data sharing, consistency, and seamless mobility

definitely what others say

Compatibility with public clouds is a growing problem. Or lack of

Monitoring, orchestration, automation of bursting, but most of all data transport concerns. The bits have to go somewhere...once again it's all about the data

DevOps and IT need to start speaking the same language. Teams need to partner up and get on the same wavelength.

Developer service availability (say database services) spanning clouds is not getting the necessary attention

Automation is one, another is understanding. Businesses have a giant junk drawer of data, figuring out what to move, when to move it, and where to is important

@yaronhaviv you mean between public and private yes?

Gaps are in cultural aspects, especially in large enterprises. Getting rid of old processes, embracing failure, taking risks not always easy.

Business models still too product-oriented. Cloud wants service-oriented revenue models.

@bpatrick001 I was thinking of how composable infrastructure is accessible via GUI, API, CLI...how that is related to cloud orchestration

Performance, public cloud su** at that

AWS is ultimate in vendor lock-in. Would be good to see some broader standards. Necessary for simple hybrid

@krishnan explain

technology gaps are hard to fill, but business model/social gaps are gating evolution.

@bpatrick001 @yaronhaviv has a good rap on the complexity of the data pipeline wrt AWS

Security is another big one. End to end monitoring and visibility (hardware up to end user experience)

@plburris can you elaborate Peter

As a board member of SNIA, we have all but begged the big clouds join and participate. They always refuse. No standards, just lock-in. We tried with CDMI, the hyperscales ignored it. Shame. Sorry to sound harsh but it's the reality of hyperscale.

@LawrenceHecht In a hybrid cloud or multi cloud scenario, dev services should be available across all the clouds. Else, it impacts both developer productivity and application performance

operational standards for app deployments. example: Private cloud: floating IP network model. Public clouds: public/private network assumed

Effective distributed middleware - distributed databases, distributed time-series databases for IoT, distributed filesystems, ability to make secure connections/agreements without central control, etc...

simple example: tech sales models still predicated on idea of intrinsic value of product, not the utility of a service.

Folks should understand that the changes aren't always big and immediate. Most of it is incremental. In the age of SaaS and web apps, people still value the human connection/support/interaction

@dfloyer Sounds trivial!

I'm reminded of minicomputer companies that told clients to bridge networks. Cisco took care of that and flattened them all with TCP/IP. Multi-cloud won't be "bridged."

Willingness to and understanding of new revenue models. Subscription models. Customer lifetime value. And how to pay sales reps in that shift.

Public cloud goes for higher level services, increase the lock-in, we must initiate standardization for services

Figuring out how to pay sales people for usage-based services is hard. If you sold hardware, it's a percentage. But when usage fluctuates and can start/stop at any time, it's hard to know what to pay sales folks.

No standards or tools for "infrastructure development." We need tooling for hybrid infrastructure design.

The role of sales and SE's isn't going away any time soon. Too much complexity still.

Not fight cloud and embrace it? :-)

Cut prices?

don't try

be invisible

The billion dollar question is whether they need a public cloud...

Biased here... but be truly committed to open standards and open source

seriously the infr guys have to get out of the way the stack is upside down now

in seriousness, think about API templates, making it easy to work with composable infra

find another line of work :)

open source will be critical wrt infrastructure co's even having opportunity to be leaders in this space

On serious note, build in security, security, security

they must move to services and platforms vs selling HW (not sell their SW assets)

@bpatrick001 Is security a "do-over" because of cloud?

They must understand that cloud instances are merely an extension of the on-premises datacenter. This concept runs through product design, dev, test, deployment, and most of all - their sales people. Infra companies must alter their comp models.

32k people went to #AWS #Reinvent to learn how not to buy hardware; this is radical change

Be willing to disrupt yourself & not try to use hybrid cloud as a vehicle to resell the same old products under a different name.

there is a lot of room for distributed cloud, my 100 mile rule (= 1 milisec latency)

Infra is required, it's about how seamless it will be (and if infra co's will make it) to move workloads between public/private environs depending on needs/costs as they flux

Focus on abstracting away the complexities, provide APIs for seamless user experience and open standards (OSS too) for portability

re-evaluate cac and ltv; big contracts do not have to be _the_ standard and customer choice will drive future revenue

@lcooney networking has moved up the stack and now networks need to be intelligent and programmable

if you cover US with 500 DCs the highest latency to a DC will be 1 milisec, higher BW (telco fiber)

HW team up with Telcos, build Metro DCs

@furrier Yes. Look at data from F5 Networks report, which backs it up

make hw seamless, easy and provide BIC security

Be willing to take risks. Focus on cultural transformation internally. Be honest about what jobs will be eliminated by automation and offer employees ability to learn new skills.

Be open to using new systems. Make tough choices to reduce/eliminate technical debt.

@yaronhaviv 5G will be hopefully a game changer

Update branding/image if applicable. (I saw recruiting commercial for GE recently that surprised me)

OpenSource is not the issue, usability & integration is, which is why public is successful (w/o Open)

Shifting from being a traditional hardware company to being a cloud service provider is a huge leap. Just shifting from hardware to software is difficult. HPE tried and seems to have given up.

OpenSource isn't the answer for everyone, with the possible exception of tools or discrete solutions (I.e., Docker Container).

I think the only way hardware providers can survive is by making it easy for their hardware to own and scale. Then it's helping the customer remove most of the middle ware that adds so much cost and overhead to the ownership.