eWEEKChat

A5: My advice - cost optimization is not a 1-time event. Orgs planning to move from on-prem to #cloud need to look at costs & optimize not only their choice of cloud platform but how best to adopt their chosen platform and continuously adapt.

A5: If a company is in the optimization phase without a solid architecture or operating model underneath, then I'd start with the arch / ops first. But assume they have the arch, team structure and operating model down ... then few things that I've seen across co

Arch - choose what u want to consume cloud-native, vs what will be layer on top

A5: Devise cloud strategy and architecture based on vendor strengths and in-house needs in your niche. Align with a true partner that is forming public partnerships with other cloud providers.

A5: I always suggest separating app-specific functionality and artifacts from core systems (monitoring, security components, IAM, etc.). Ensure consistency across core systems and then let app-specific choices be made to achieve specific functionality

Team - invest in a cloud center of excellence COE for good X-functional decision making.

Ops - have a good mix of domain expertise and programmability. If you strive too far right or left, you'll have a mess.

A5: If your use cases warrant multicloud, embrace it. Optimize for the business outcome rather than platform feature / function type tradeoffs.

(edited)

A5 Example of app-specific functionality is database choice. I don't think there's any advantage in a company operating its own, say, MySQL implementation rather than letting an app use a managed MySQL service. More important is ensuring sec reqs for database imp

A5: In a #datacenter world it was all about one-time or periodic #Capex optimization. With #multicloud, it really becomes a challenge of continuous Opex optimization - so you're not surprised by that cloud bill at the end of the month.

@BMCSoftware Spot on - The "continuous" part of this is the hard part. Doing this one time is great but waking up next day to see a completely different set of optimizations screws this up.

A5 @bernardgolden great point. Don't use least common denominator thinking to stay "cloud neutral", its about being "cloud leveraged" to help your business get the best outcome

(edited)

A5 reskill your people, ruthlessly standardize, set policy, enable the business to succeed.

A5 Establish the right metrics so you can tell if multicloud is the right option for your company, and if your implementation is actually yielding the results you are after.

A4: Proactive market standardization, building with open-source, built-in security, and security vendors as secondary measure

A4: Adopting tools that accurately map the business service across the #stack, ingest a large volume, wide-variety of #data (metrics, events, logs) from multiple sources & use #AI-powered intelligence, #ML & analytics to proactively prevent & quickly fix issues.

@Chris_Ehrlich completely agree, open is the best way to avoid trading off one type of lock-in for another

First deep breaths always help :-). But in all seriousness, invest in a layer on top of cloud-native capabilities. Think about how to achieve outcomes / experience / simplicity. We have several examples of this - Terraform for IaC, Redis / snowflake for data/DB

A4: Also, specialist vendors will solve the central management issue. The opportunity to own the category is too large to leave unsolved.

(edited)

@ramesh Is this top layer a management layer?

@Chris_Ehrlich +1 for using opensource and strive for some level of standardization. Challenge is knowing when to use cloud-native (because of its advantages) vs when to invest on top

A4. there's a serious resistence to change that is preventing orgs from capitalizing on new technology.

Management & visibility certainly, but it cloud also be a shim on top of core capabilities. In the infrastructure space, this could mean security APIs, proxies etc. Ultimately you want to operating on top of this common layer - not the underlying constructs

@ramsba When adding a top layer across different cloud environments is important and key is mapping that to business services.

@BMCSoftware 100%

you can always hire a bunch of consultants. :-) But seriously, orgs need to take responsibility and establish a Dev_Sec_Fin_Ops approach with consistent policy, security and oversight.

@cloudDay_2 critically important to realize that the vendors, skills, tools, workflows that worked for you in the past, don't necessarily apply to the future.

A3. Probably the single biggest I have comr across is consistency across environments. If you draw a matrix of capabilities (vs) hyperscaler - you'll see holes and inconsistencies that will make your head hurt. Each service - compute, network, data will have nuances

A3: Complexity involved when using vendors that have different management, monitoring, security, and service offerings. It can quickly become a combinatorial challenge to keep apps in multiple clouds operating properly, along with skill breadth issues

A3: Interoperability, central management, and security as the multicloud universe expands

A3: One challenge we see is getting visibility into the entire application / service stack, particularly large enterprises who run on-prem, #cloud & #mainframe – where their business services span #multicloud (&datacenter) environments.

...and trying to achieve consistency with an inconsistent underlying set of capabilities is not easy

@bernardgolden ultimately its about ensuring a consistent experience across multicloud platforms.

A3 Multicloud does not mean least common denominator, providing a consistent user experience while taking advantage of the underlying cloud provider's capabilities is the balancing act.

A3 This is where open really helps IMO. Open APIs, Open protocols and formats can really help drive consistency cleanly

@ramvzz So open source is the holy grail of multi cloud computing?

A3: Open source is a way of achieving consistent functionality and operations across clouds; however, it comes with the cost of taking on curation, integration, and operation of the user's open source components.

org model. ITops is not the way.

@cloudDay_2 (liebow has joined the chat -- technical problems)

@cloudDay_2 What's wrong with ITops in this context?

A3 I find many orgs start down the path of DIY systems based on open source, but find that it requires more investment and ongoing involvement than first assessed.

Open source has its challenges too, but in this context, its powerful, especially compared to the alternative

treating the cloud like just another data center is not the model. You'll hit a wall with antiquated processes and tools.

A2: The simple answer: cost & functionality. Every #cloud provider is looking to differentiate themselves from each other - evolving from purely #Infrastructure as a Service to delivering an extensive Platform as a Services with a broad range of managed services.

A2: Ongoing cloud migrations, cloud competition, vendor lock-in, domain-specific solutions, and redundancy as cyber threats increase

A2. If you focus on whats driving this from within the Enterprise, really 3

A2: Its' customer needs. Optionality is underrated except when you need it. With evolving cost and risk concerns, customers are looking for solutions that best fit their needs, be it on another cloud than their primary choice or even onprem.

1) Multiple business units with each BU having autonomy to make cloud and devops decisions 2) M&A - when a company acquires another 3) app centric development - where some workloads run better on 1 cloud vs other

@ramvzz Love this: " Optionality is underrated except when you need it."

@ramvzz +1 on customer needs. choice and optionality rule this decision in a staggering way

A2: and, as @ramsba pointed out, inevitably a company acquires another which has made a different cloud provider choice and, whoops, you're multi-cloud!

@bernardgolden spot on. one day you are consciously single cloud, one acquisition later boom - you have to deal with multicloud complexity

@bernardgolden "Acqui-Cloud" :)

A2: One topic that has been mentioned so far - regulatory compliance. Customers sometimes have to consider alternative clouds to stay compliant with emerging regulations regarding data governance. For them multicloud is existential.

A2 Heh

@ramvzz Yes, have frequently seen that with financial firms.