Cloud Workload Repatriation with Deloitte’s David Linthicum
Transcript
all right welcome to another episode of the CTO advisor podcast we have a special guest online I don't get to talk to many other folks from the management consulting arm of our industry the big four who do this really high level Consulting it's it's technical in nature but we're talking about large strategic problems I have with me my pleasure Dave lithicum Chief Cloud strategy officer at Deloitte Consulting Dave welcome to the podcast well thank you very much thanks for having me on I appreciate it so
Dave we we were reminiscing at least I was reminiscing you're in the Thrills of management consulting today where you get to where you used to get to go to all kind of far Plum places of the world China Australia I'm assuming Parts you know every part of the U.S to help folks with their strategy around I.T technology today we're going to talk about repatriation but how's the travel these days not as much but is coming back obviously the airlines aren't cooperating in some instances we have
to get around those snafus but uh you know Consulting is you know always going to be a traveling you know traveling profession so it's coming back slowly but it's going to be coming back so we've had plenty of conversations over the past few years in the industry around repatriation we're not going to debate like is there repatriation or uh should you repatriate or the level of activity the simple fact is that there's a use case for it and let's talk about first uh you being a
Management Consultant let's talk about the business value what are some of the considerations uh that's driving repatriation yeah I think it comes down to looking at a workload and a data sets ability to bring value back to the business on the cloud versus on a traditional system or even on a private cloud and looking at what those metrics are and so what happened during and certainly this happened during the pandemic is people had a Mad Dash into the clouds in other words they moved lots of
workloads lots of lifting and shifting went on lots of data you know data set movements things like that you know now they're getting the bills and now that we have a lot of fin Ops in place and cost observability they're seeing the fact that this is a lot more expensive than it was on my traditional system so let's take a look at this and reevaluate if I'm going to have a better deal or bring more value back to the business by moving it back to a
similar system hopefully operating them cheaper and usually the hardware and software is going to be cheaper or leaving it in the cloud and even looking at the soft benefits of agility and flexibility and all the things that all the abilities that come along with the cloud um you know versus some of the limitation that's going to be there and moving it back on premise that we know having dealt with the cloud for the last 15 years you know what those are but in some instances not
all it's going to make more economic sense to put this workload or put this data set back on a more traditional system in a traditional data center in traditional operations than it is to leave them in the cloud and those are the kind of uh problem domains that we're seeing right now so I know it's you know it's hard to have a conversation with two people that's historically been on the management consulting side because more than any other part of the industry the the answer is
really strongly it depends it depends we'll we'll we'll try to resist our natural nature to answer it depends what what are the what are some of the characteristics of workloads that are appealing to people to bring back on-prem yeah that's a great question uh so it's going to be inter workloads that typically aren't localized for cloud native features in other words they're not bound to the security system on the public cloud provider they're not dependent on uh the scaling system they're not dependent on particular apis
and so if you put a lot of money in time into building those kinds of systems or even refactoring applications that move to the cloud those are going to have to be unfactored to put them back on premise so you have to kind of and so in many instances by the time you do that I think you're seeing the value and running those applications on the cloud and data sets on the cloud as it is so those would be an instance of where it would uh
you know probably make not make sense for it to move back in the cloud um some things like if you're doing the same thing over and over again in other words you have a common system that just does the same typical patterns of processing there's no dependency or no benefits in running on the cloud then those may be a candidate repatriation large amounts of data storage you're doing large amounts of unstructured data storage on the on a public cloud provider and you're getting a huge bill
from that certainly if you're doing a lot of egress and Ingress in the cloud then putting it back on premise may make good economic sense the fact you could be paying 20 of what you're paying for in the cloud and so um yeah avoiding the it depends thing you know it's really just kind of looking at this from a business metrics perspective and not and kind of taking off your technical hat if you're going to debate um you know open systems and Cloud systems and Agility
and those sorts of things versus on-premise systems that's not the argument you're looking to have you're looking to say well we exist in I.T to return value back to the business how best can we do that in some instances it's going to be moving it back on premise where it's going to have more economic value so let's say that this we've made the financial decision that we're going to move back on premises one of the things that I've read and I've heard and I've talked to
people is that they've lost the skill set they've lost this skill to manage on-prem systems matter of fact some people never had it repatriation isn't just about coming back on Prem but going on Prem for the first time there are companies that have been born in the cloud that they've operated only in the cloud what are some of the kind of high level considerations about restarting your on-premises operations should I go out there and look for a place to break down and build a data center
yeah I wouldn't put that on anybody um I I think really you consider that in the equation and so in other words if I don't have the skills on skills in-house that are able to maintain these on-premise systems then I have to look at what it's going to cost to obtain those skills and if you know and there's actually a diminishing number of people who who do that stuff now everybody has clouds as reassessed and retrained in the cloud so that could be a consideration that
may be a reason not to repatriate so in other words okay we're going to save uh you know 20 30 percent on these workloads and these data sets moving on back on premise you know all in operating costs things like that but if we can't get the skills you just you just don't do it and by the way people are making those decisions now they're even they're making architectural decisions not based on what technology is going to be best a breed for the particular problem they're
looking to solve but you know around what skills they're able to obtain and maintain and and so that's going to be a core decision and I think that's very important it's probably a reason a lot of people aren't going to repatriate because they don't have the ability to maintain those workloads even though they're going to be cheaper to run so you know this is something I think you're uniquely positioned to answer as Architects we can look at these things and say oh this is an apple
it should go on the apple cart this is uh orange it should go into the band that I used to make orange juice simple decision when the land is turned over and we say oh it's time to build to to grow squash here or whatever the case as technical people we can look at a situation and make that decision politically if I sold the idea that three years ago we're going all in on cloud and the landscape changes how do I begin to have that conversation
with the c-suite that yes three years ago that was the best situation but we need to start to tap the brakes on our cloud roll out and think about keeping some of these data center assets or whatever the current infrastructure yeah it's all data driven and so I would go to the board of directors whoever this you know c-suite was hopefully you didn't make the decision and you know declared uh you know declared everything's going in the cloud that's obviously something you have to kind of
look at as a wise choice you have to look at everything goes into the proper Cloud platform whether it's cloud or not and say we have new metrics um this is costing us more because of this this and this and the reality is you know the cloud environments from a cost perspective or a changing Dynamic and so in other words their prices are going to go up and renegotiating Enterprise contracts things like that and just like we have Outsourcing and we have people delivering power and
water into our houses and businesses they may become less economically desirable if they change their thing so I would go there and say listen look based on these metrics we may have an opportunity to save some money by moving some of the workloads and it's typically going to be some all and the data sets you know back on premise are moving it into a managed service provider or a Colo provider that's typically what people are doing now they're renting the space and we can save say
a hundred thousand dollars a month if we make this particular particular move now if they come back and they say well you said we're all in Cloud you're contradicting yourself that's a problem and so in many instances you got to have the political will to really kind of make the assessment as to um I wouldn't say a mistake was made but we need to look at this in a much more pragmatic way than we did five years ago and reevaluate some of the workloads at the
end of the day it's kind of right sizing and it's tuning because we're not taking you know the 500 workloads and the uh you know 20 petabytes of data that sits in the cloud and then relocating it back on premise we're taking aspects of that in other words we're taking analytical database here is unstructured storage file system storage archival storage you know these things are kind of boring systems that will run perfectly well on premise as they do in the cloud and just making a decision
that we're going to save money by pushing them back on premise and I would have a probably hopefully a compelling argument to make to the boards of directors and CEOs and hopefully I you know didn't make the decision or didn't declare that we're all in on cloud that's something I wouldn't do personally but it comes down to a business decision just like moving your business from one building to another from one city city and state to another city and state based on tax advantages you know
those kinds of business decisions are made all the time it's really kind of no different yeah some people call it the heel to die on Etc and I think this is a great example of our audience has a crossover there's kind of the Enterprise architect looking to expire to become the CTO one day and this is one of those political decisions that you have to get really good at like is this worth spending political capital on and when did you find that a decision like this
is worth spending political Capital I'll hearken back to an example that my audience is aware of I had to make the from an Enterprise architecture perspective I had to make the political decision that we're not going to move from HP ux to Linux uh for our sap instance the business had just gone through a really painful consolidation of sap and I would have had to ask our CFO to spend more political Capital to go to hpux from from hbx to Linux at the end of the
day I had to do the assessment do I want to spam my political capital on that transaction where does this fall along the this conversation line yeah I don't think it's any different than in this scenario you're just described it's a great example it's a great analogy and I think the reality is you have to make a decision as to whether it's worth um you to make this assertion which by the way they won't expect they're not expecting that you make this assertion and doing your
duty as someone who's uh fiscally responsible responsible for the business and so that kind of depends on the person you are and who's making making the thing in many instances these kind of audits and these kind of Assessments around repatriation are going to be sponsored by the sea levels and by the boards of directors those are kind of looking for an audit kind of to make sure that they're investing in the right platforms and that we're changing the platforms they're optimizing them to be better at
it so I don't think there's a lot of political law loss that can happen there depending on the organization and so what you're doing is just saying that we're going to make some minor adjustments because that's what you're doing you're moving aspects of systems we're not doing a huge repatriation effort to save some money but if people are going to uh you know view that as something that's a negative that's something you have to consider you have to consider the culture of the organization there's some
organizations I worked with I wouldn't have no problem you know going into the CEO's office and telling them we have to go ahead and make this change even though it'd be confusing to them uh in some instances others I wouldn't because I would be some really fired and so you have to you kind of have to make that decision unto yourself and even for people in the organization am I going to you know maintain uh you know their ability to work here and have a good
you know have a good working relationship with the with the company if they're part of the team that's uh suggesting this stuff and so it's not not just you but you have to make you kind of have to use your head when you make those decisions my my bet on this this you know at my Advanced age is you're there to return Equity back your return value back to the investors you know people who own the company however you want to do that even if you're
admitting to a mistake I I think that's worthy I think that's just doing the right thing uh some people may disagree with that so you put the data and the information in front of the people who are making the decision recommend a course of action they can either say yes or no but in doing so you're doing the right thing to the investors the people who actually own the company well that's the fiduciary responsibility at its essence so this is one of those things you can't
get a certification for this is just experience uh this is uh now that I have 25 years of Industry experience I can look back at what my manager told me five years in Keith that it takes 25 years to get 25 years of experience I now better understand that last question this is one is just kind of a curveball for you but I was going over my response for call for papers and I was surprised to see that you submitted a response to the call for
papers there's this term supercloud meta Cloud distributed Cloud going on and you're you've submitted a paper to talk about that preview that concept for us yeah it's it's fairly simplistic if you think about it we're trying to remove uh hybrid clouds and multi-clouds are are complex and that's where people are running into the problems and operating them so in other words they become there's you know you go from a uh you know a thousand Services under management to 3 000 Services under management and by the
way we're not giving any more money to operate those systems you can't hire more people can't build more tools things like that so we have to figure some way to mediate that complexity and one thing you can do you can certainly you know try to underwhelm the operations and end up making mistakes and make having outages and things like that or think about more creative ways to do that so the idea with meta cloud and super cloud whatever you want to call it is to build
this Loosely coupled cross-cloud uh layer or hybrid across hybrid Cloud layer however you want to do it that is able to provide a central way to deal with security and operations and governance and data storage and data management so we're not doing it in different ways on each individual system you know hybrid clouding would be doing on the private cloud and the public Cloud are doing on the traditional system in the public Cloud multi-cloud are we doing on five different public clouds in different ways we're
implementing security five different times and data management five different times and that's where the complexity comes in we just can't keep enough people around to kind of maintain all of these moving parts to make it happen so the idea and by the way it's an older architectural concept the idea with uh meta Cloud super cloud whatever you want to call it is having this Lucy coupled layer that sits above the clouds by the way they may run on particular Cloud providers but they're able to extend
and reach out and communicate they logically sit above the cloud providers where we're trying to consolidate lots of different things into a single layer thus reducing the complexity thusing allowing thus allowing us to operate these systems at scale and thus allowing hopefully to reduce the money it's going to cost to operate these things at least that's the idea and then uh as the outro so to speak for people that are not familiar with your work you have a book coming out can you tell us about
the book and when it will be released yeah in April it's called Insider's Guide to cloud computing does discuss a lot of the topics we talked about on the show but basically it's looking at cloud in a pragmatic way so folks who are you know trying to figure out how to make the right decisions uh can kind of get an Insider's view because you know CTO I worked on the product side and certainly the manage the Consulting side and uh you know uh all over the
place and different Technologies over time as to what they should be thinking about and also some of the things I don't think other people are telling them certainly we're not hearing it out there from the vendors uh in the technology Community as much so it's getting to uh what you should be thinking and how to think differently about using cloud cloud providers and using them effectively Dave I really appreciate you taking time out of your schedule to come and talk to us about Cloud repatriation not
from that normal should or shouldn't we but really detailed data if people want to find out more about the CTO advisor you can find us on the web the ctoadvisor.com you can follow me on Twitter at ctoadvisor all of David's contact information will be in the show notes so you can find that on the ctoadvisor.com until the next episode of the CTO advisor I haven't said this in a while share the podcast with your wife your husband your grandma your mother put it in their pie
catcher they will be pleased to know more about Cloud repatriation talk to you next episode
