LocalData

When data can’t move offsite
Community discussion and strategies for managing data can’t or won’t move to the public cloud.
   3 years ago
#localdataSpawning a new hybrid cloudBeyond just moving legacy apps to the public cloud what about bringing cloud services to #LocalData?
John Furrier
I just got DM for a question I'll share here: What is the inhibitor for moving data to the cloud. Seems so obvious?
Bryan Champagne
compliance is the biggest hurdle that we experience
Stuart Miniman
from a physics standpoint, moving data is always difficult. It's all about your applications, and where the data is needed.
John Furrier
IMHO I say Fear
Christopher Jones
In some cases ingestion is a big problem. Massive scale images, video, genomics, require massive amounts of data to be moved. some of this movement is very difficult to do - sometimes it’s even locally demanding!
Aaron Cardenas
Security and time to retrieve
John Furrier
While we are on this topic let me add: What % of workloads do you see going to cloud? What apps or data have move first?
Sagi Brody
This is where edge connectivity and data centers can play a big role
Sagi Brody
When we talk about DRaaS and BaaS, you have to consider the speed of recovery/copy. If you're working with a local/edge location you can always drive over and pick up your data. This sounds low tech, but CIOs love this.
Robert Hout
I think its a data gravity issue as much as anything. Where are the apps? The data will follow. Unless of course you have compliance issues.
Christopher Jones
A few workloads generate most data. An enterprise with hundreds of apps may have fewer than a dozen generating 90+% of data.
John Furrier
I always here low hanging fruit is testdev
Sagi Brody
I believe a big hurdle is the pre-existing network security an organization has already defined. Being able to consume cloud services w/o traversing the Internet will allow for much more common adoption.
John Furrier
here's a huge amount of enterprise net-new application development on top of the cloud
John Furrier
after dev test there a a few others nne is digitization. Almost all of these big enterprises now are going through IT modernization where they're basically digitizing all their assets
Sagi Brody
we're starting to see ERP systems now moving to the cloud in addition to the typical stuff such as webapps..
John Furrier
then I see analytics. It's just never been easier and less expensive to collect, store, analyze, and share data than it is today with the cloud
Sagi Brody
for example #SAP in the #Cloud is starting to become very common. I would consider that one of the outliers..
I am John White
fear, location physics, capital rich business all inhibit an OPEX model whether it is in a customer or provider DC
John Furrier
so much easier and so much more cost-effective to run analytics in the cloud
John Furrier
and then mobile because most companies have employees and customers that now have three screens between laptop or desktop and tablet and phone, and most of the applications weren't built with those other interfaces in mind
John Furrier
after those low hanging fruit areas that migrate to the cloud then companies will move their entire datacenters over
John Furrier
so workload moving to cloud summary: 1) test dev 2) digitization of entire business 3) entire datacenter footprints
Eric Ransden
I still see some customers that see it as a threat to their job. They just don't see the future of how they can evolve to better serve the company.
Rob Peglar
John you are right in many cases because the ingest is done directly into cloud resources, and computed upon. But in significant cases this process is much more cost-effectively done on-premises - it's very volume-dependent. One size does not fit all.
I am John White
It is cost-effective if you don't have predictability in your workload. If you do and the environment is relatively large(PB or more), it might be better to capitalize and then burst when needed,
John Furrier
@dahowlett said @furrier inhibitor? Multiple but - don't even think this is simple stuff - it aint.
ConsumingML
Compliance and security. Orgs like FSS and Consultants can have unlimited liability. If your cloud agreement is limited, that risk is now borne by org.
ConsumingML
customers not so much worried about data in motion or transient as they are data at rest in cloud. If you can hook analytic cloud app to on-prem data, sweet spot right now.
Rob Peglar
Good point John about burst. This can be an efficient method to use cloud resources. Rent a bit of time and space only when you need it - helps to solve the Erlang problem. But one must consider transport (network) costs when doing so.
Steven Hill
The security problem only gets trickier as data begins to exist in multiple locations. Not only from the perspective on on-prem or off, but will eventually spread to WHERE the data will be allowed to reside
John Furrier
Agree Steven: Perimeter #security is dead and #cloud has to solve this and #IoT only makes it more complicated
ConsumingML
think we are already there Steven. European customers can't have data in NA data centers. Becoming a very large issue. If you don't DCs, data locales and apps in EU, you can't do biz in Europe
John Furrier
Q10: What does the group think about Kubernetes movement? Tempest in teapot or real?
Sagi Brody
Its real. But do and get "shiny object syndrome". Are you running use-cases that can utilize thousands of stateless instances to achieve a goal? No? Are you interested in building one or rewriting your app? Awesome, its perfect!!
John Furrier
Did you see @cmcluck new company got $8.5m in funding http://siliconangle....
Christopher Jones
Docker with teeth?
I am John White
Very real and very confusing for the average IT person. Unless you are starting a new or completely rewriting your monolith app you are going to struggle to find value.
Stephen Pao
- Containers are real. We're a consumer, as we've containerized our whole stack, and we'll also be a provider (more on this later!)
Sagi Brody
What happened was everyone jumped on the bandwagon though, and people were trying to fit a square peg into a round hole... Again its all about the use-case.. But if you go back to "throwing away data" and real time analytics.. Great use- case for it.
John Furrier
I think that Google is trying to level the playing field with respect to Docker and AWS; Why wouldn't AWS just go native Kubernetes
John Furrier
Some are saying the open core license strategy won't work anymore bc mgt layer will be open with Kubernetes
John Furrier
I had a great conversation with Microsoft team at #Kubecon Brendon Burns is totally into opening things up
Sagi Brody
It has a ways to go before it hits the Enterprise, but it will happen
Sagi Brody
One technology to track which may be analogous is OpenStack though, there's been some highs and lows regarding its adoption and investments..
John Furrier
@webairsagi I see the same thing but the nice thing about Kube is that it has primatives above the PaaS layer
John Furrier
@webairsagi It seems to be forcing the PaaS to be thinner which makes it more lightweight always a good concept in cloud
I am John White
@webairsagi I think Docker will hit the enterprise first as they are creating an ecosystem around bringing your legacy IT practices and apps to containers where k8s is focused on new app dev.
Stephen Pao
At #KubeCon in Seattle, we were seeing lots of enterprises there. Thought it might be a "vendor show" but we were surprised!
Aaron Cardenas
Assuming P2V was a 20-1 net gain, V2Container is a 200-1 gain. Believe the hype
Sagi Brody
@steve_pao part of that is shiny object syndrome and people not wanting to get "left behind". Its alot less common to find folks running them in production
Steven Hill
Docker got an amazing amount of buy-in across the industry really quickly, even though it still needed lots of work.
John Furrier
@steve_pao its being expanded to CNCF cloud native computing foundation is running it like the IETF
Stephen Pao
@webairsagi I agree with the point. Still early days, as 2nd Watch and 451 stats show AWS usage in elastic containers is still not high like EC2/S3.
John Furrier
I'm always skeptical of vendors "moving the goal posts" to change the game in their favor; so to me community and users drive the relevance algo
Steven Hill
But I think it's experiencing some of the same challenges that OpenStack has when it comes to enterprise adoption. It's an awesome framework with a ton of potential and industry support, but it's incomplete and requires hardening for enterprise us
Bryan Champagne
enterprises want a better way to do things and are exploring alternate solutions but the production adoptions have been slow so far
Sagi Brody
There's a huge increase in complexity, management overhead, and security.. It has to be worth the gains. The common management platforms used to handle these tasks for enterprise are still not there yet for containers.