ATR2500x-USB Microphone & Logitech BRIO: Have you ever thought about how many

copies there are of your production data?

There's the primary copies, snapshots copied to another array.

Cloud snapshot stored in object storage.

A copy each for dev and test other copies for analytics and of course, dozens to

hundreds of copies for backup and Dr.

Ever thought about who has access to all those copies, how long they're going

to stay around and what they're costing the company while they hang around?

No.

Well, your welcome.

You've just entered the world of copy data management.

It's perhaps the most boring and the most important thing you can do to

protect your company's information.

And save money at the same time.

Hi, I'm W.

Curtis Preston aKA Mister backup.

And each week on this podcast, we dive deep on one topic

somehow related to backup Dr.

And ransomware.

We turn unappreciated, backup admins and to cyber recovery heroes.

This is the backup wrap up.

W. Curtis Preston: Hi, and welcome to the backup wrap up.

I'm your host, W.

Curtis Preston, I have with me my LinkedIn algorithm commiserater

Prasanna Malaiyandi, how's it going?

Prasanna.

I'm good, Curtis, and I am sorry that your LinkedIn

stuff did not go as well as you were hoping it would go, but I think it's

because they've probably caught up with you and realized how you use

things and are like, okay, Curtis, Mr.

Backup, we're going to change the algorithm just for him.

W. Curtis Preston: Yeah.

I don't, I don't think so.

I think, you know, it's the, you know, when, when things go wrong, we're

taught, what did you last change?

And when I was taught, like when I had, we had these consultants

that taught us how to use LinkedIn.

One of the things they said was, don't post within 24 hours.

Right.

Don't, don't bang.

You know that, that wait at least 24 hours before you do your big post.

I didn't wait at least 24 hours.

I tried something different yesterday where I posted to a group, um, with

the link and, and just to see what kind of in attention I got there.

And then today I posted to, you know, the greater LinkedIn.

And, um, it was not 24 hours, it was more like 18.

And so I think that in the end was my, my boo booo.

So I won't make that boo booo again.

Um, and uh, but yeah, I, yeah, so I think there are reasons.

I don't think they just changed the algorithm

I don't know, maybe it's just Yep.

Curtis, disable him.

Done.

W. Curtis Preston: yeah, I was having too much success on LinkedIn.

All right.

Well let's get onto the news of the week.

First story of this week is a follow on from a story from a week or so ago,

and that is what we originally referred to as the one password slash Okta hack

really turned out to be an Okta hack.

You want to talk about it Prasanna.

Yeah, so this is the one that we've been following along.

Um, Okta had someone breach their environment, which then

let them get into one password.

Um, turns out that Okta recently finished their investigation and

they published their findings.

The findings were that one of their employees.

Had a corporate device, they were using it to access a service account at Okta.

Right.

Typical.

That's what you would do, right?

They needed access to it and then they used their Chrome browser and

they logged into their personal email account using the Chrome browser, and

then when they went to go access using that service account, they said, save

password, and the password for the service account then got stored in.

The Chrome browser's password manager associated with their personal email

account, and then supposedly somehow the employees, either personal

device or their Gmail account was hacked, and that's how the bad actors

got access to the service account.

W. Curtis Preston: Yeah.

So much to unpack here, right?

Um, first, you know, just to understand, you know, we've talked about the Chrome

and password manager before, and I, you know, we've put it in the category

of it's better than nothing maybe, um, because of this problem, right?

So it's better than nothing in that.

It would allow you to have a unique password for every

site, which that is good.

But the problem that I have with the Chrome Password Manager is that

it stores the data in such a way that you can pull it out, right?

When you install one password or um, you know, Dashlane or any of

these other guys, they will pull.

Uh, they can pull your password stored in your browser out and then put them

into your new password manager, which sounds really nice and convenient.

What it should tell you is that there is an API to pull out the plain text data.

All you have to do is ask the browser apparently.

There was another article that was written in our Technica that basically said, no,

Okta Senior Management, not an errand employee, caused you to get hacked.

Do you wanna talk about that?

Yeah, and the big thing here is.

ARS Technica article talks about the fact that there should have been IT

policies in place to prevent these sort of things, and they covered the two

that I remember off the top of my head.

One is that for a service account, you shouldn't just allow anyone who has a

credentials to be able to access it.

You want to be able to limit the ips that can access it or do other things like

that to make sure that you're restricting access, especially a service account is.

Pretty powerful.

And usually you don't have MFA associated with it because it's used

in cases of running automated scripts or automated access, and so you

can't really do MFA in those cases.

So it makes sense to have the service account, but at the same time, you should

restrict who has access and their ability and where it can be accessed from.

The other part that they also mentioned in the article is that Okta, their

IT policy should not allow personal.

Accounts to be logged into from like a web browser to avoid the same issue.

And there are multiple ways you could do this.

You could either have a web proxy, you can, using Google

Workspaces, you can actually restrict what domains are allowed.

And so there are multiple tools that they could have done, but they

did not have a policy in place, and that's what led to this issue.

W. Curtis Preston: Yeah, I'd say it's gotta start with the policy, right?

Don't log into your personal accounts from your computer, um, and then,

and then, you know, do what you can to use technology to, to stop that.

I also wonder if, if there's a way through, through technology,

if we could go and blow away any stored passwords in any browsers.

I wonder if that's possible

I'm sure with the device management software, there's

W. Curtis Preston: Yeah, the di device management software.

Yeah.

Yeah.

Okay.

All right.

Well let's talk about some good news.

Um, and you know, I first wanna just put out a disclaimer.

We don't purposefully go looking for Druva News because you and I used to work there.

We just Google the same stuff that anybody else does, and it just so

happened that in a couple of weeks, Dr.

Made.

A couple of, uh, big stories, and I think this one is probably my

favorite story coming out of DVA in a while, and that is that they now

support natively Azure backup of VMs.

And specifically, not only do they support basically the, the snapshot.

Method, which is the supported way to back up VMs inside Azure.

They are then able to export those, uh backups, de-dupe them,

and store them in the Druva cloud, it's what they do with AWS VMs.

You get the best of both worlds.

You get that native backup and recovery, and you get the cost

savings of, um, the deduplication.

, Prasanna Malaiyandi: I think for some of our listeners who may not be familiar.

Why is it important, that second part that you mentioned of how they work,

that they're taking the data, the backups out of the customer account

and moving it to the DVA account.

Why is that so critical in your opinion?

W. Curtis Preston: Yeah, I think that, yeah, thanks for asking.

So I would say that it's because of, you know, we're, we're always talking

about getting an air gap, getting something that, that mimics an air gap

if, if, of all your backups are in your.

Uh, AWS or your Azure or your GCP account, then that account gets hacked.

They take your backups with them, right?

There are way too many stories about this that, uh, you need to

get as much separation as you can.

One of the ways to do that is to put it into another account that you own.

I think the best way to do it is to put it into an account that you don't own.

Right.

Uh, in this case, this is what Druva is providing for both AWS and Azure is that

it's pulling the data out, de-duping it stored in there, and by de-duping it,

that's where they get the cost reduction.

They're saying that it results in an overall reduction of TCO of 40%, which

is, uh, you know, a solid number.

Right.

Um, and there is a cost to pulling the data outta the account.

Right.

I know that was the first thing you asked me was what about the, the,

uh, the Egress cost, right?

Yeah.

Uh, but that cost is clearly offset by.

The deduplication features, right?

So you get the, you get the best of both worlds.

Um, and this is good news for.

Druva, you know, they, they acquired Cloud Ranger like five years ago and

I think five years minus one day ago.

I said, okay, that's great.

What about Azure?

Yeah.

W. Curtis Preston: And uh, so I'm glad to see that they're finally supporting

it and I'm assuming that GCP is next well that is the news of the week . Alright.

On this episode of our continuing backup to basic series, which

of course what we're doing is we're working our way through.

Modern data protection.

My latest book here, I'll show, hold it up for the camera for the

27 of you that are watching in the, the video version of this.

We have a, we have a nice following on the audio side, but you know,

I, I just think nobody knows.

You can also watch this on YouTube.

If you go search on the backup wrap up on YouTube, you can, you

can watch our lovely little mugs.

Um, I put a lot of effort into that video version,

Oh, I know you do.

W. Curtis Preston: you know, nobody, nobody watches it.

Just someday maybe we will, we'll be discovered on YouTube and then,

We take off.

W. Curtis Preston: our, our, it'll just take off.

But anyway, you know, you might think, and I did not do this on purpose,

but the YouTube copy might just be, it's just one of the many copies

of this show that I need to manage.

You see what I did there?

I like what you did there.

W. Curtis Preston: So we're talking this week about this phrase that I

think, I think at one point it was really big, and then like lately, I, I,

I don't hear the phrase too much, but again, it's a backup to basic series.

So I want our listeners to know what this phrase is when it comes

up in conversation and, and to know.

It's pros and cons, uh, and the ways in which it might manifest itself.

And that is, of course, copy data management or C uh, dmm.

CDMI never hear I hear any, I never hear anybody call it CDM,

I, I haven't really heard.

Yeah.

W. Curtis Preston: They

always, they always seem to say copy data management,

Copy data management, copy data

management, copy data management.

W. Curtis Preston: copy data.

It rolls off the tongue, you know,

W. Curtis Preston: of copy data.

Does it roll off the tongue?

it does.

W. Curtis Preston: Um, so what do, do you want to sort of give the basic

concept behind copy data management?

Yeah, so copy data management is really.

It literally is what it says, right?

It is.

How do you manage the copies of data in your environment?

Uh, it, depending on what you want to think about backup

is one copy of the data.

I know we've talked about replication and snapshots and CDP and near CDP, right?

Those all create copies.

Your archive, if you go off.

Do archiving, right?

That's another copy of the data.

And so copy data management is sort of how do you manage all these copies and

the life cycle of those copies because it's not just, okay, where are the copies?

How do I create those copies?

But also I.

How long do I keep the copy around?

Because different copies in different places will be kept

for different periods of time.

Snapshots, maybe I'm only keeping for 14 days backups, so maybe I'm keeping for 90

days long-term retention slash archiving, maybe I'm keeping those for like 10 years.

So how do I manage the retention process for those copies?

And then the last bit of copy data management is really around how can

I now start to use those copies?

Which I think is kind of what a lot of people have now started to think

about is, okay, I have copy of the data sitting there and I wanna just

leave it sitting there because it could be useful for other business

purposes for me to drive insights from.

So how do I now start to manage the lifecycle of using a copy, destroying a

copy, and keeping that around if need be?

W. Curtis Preston: I Would describe it as like perhaps a more holistic view.

Of all of the copies that are out there.

You know, in, in our world, we tend to worry primarily

about one of the copies, right?

Or maybe two of the copies, right?

We want at least two copies, right?

You know, we wanna, we want a backup copy, hopefully on-Prem, and

we wanna backup up copy off-Prem.

So that's kind of two copies.

And then a lot of people primarily focus on the, on the, on the,

on the primary copy, the, you know, the, the production copy.

Um, the, and you talked about copies that might be kept for archival purposes.

I think the one, the only one that I, that I, uh, thought that you left

out was the, the development side of

Test and dev.

W. Curtis Preston: because that is, yeah, yeah.

That is another big area where if you are a, if you're the type

of company that does any kind of development or any kind of testing.

Where you're, you know, where you, you want to have this other copy of your data.

What you often want is you want production data, right?

You want, um, a, a copy of your production data to use for testing you.

Well, I'm just saying if you want, I know, I know why you're

wincing, but, but if you want to,

Prasanna Malaiyandi: Before it goes out to prod.

I agree.

Yeah.

W. Curtis Preston: Right.

If you want to test it.

Right.

I understand the, the concern that you're, that you're talking about.

Um, but, and we, and we can talk about that, what, you know, that that's one of

the things that you deal with in a copy data management, uh, configuration, right.

The, the, the closer that you can use sandboxed.

Information.

Right.

Um, so with, so for example, with Salesforce, right, you can very

easily create a sandbox environment of your production environment.

And it's got all of your production data over in that other, um, world, uh, so

that you can test whatever new thing that you want to do without, uh, destroying

your, yeah, without impacting your

I, I, I go back to the story that I think

you've told us before on the podcast about how you, what was it?

You downloaded Salesforce, you exported the Salesforce records,

you sort of screwed up the changes and the rearranging of the tables

completely ruined all the records, and you luckily had a backup, right?

And you were able to go

W. Curtis Preston: Yeah.

I, I, um.

Yeah, so first off, Excel, whoever wrote Excel and made it way too easy

to sort just one column, why would anyone in the history of people want to

sort just one column in a spreadsheet?

It's got to be the non-fat, right?

It's gotta be the one out of a hundred use case I, for the life of me.

Can't think of.

Any use case where that would be the case.

But, uh, I sorted the phone number column, you know, and, and, and

only the phone number column.

And so I put all the wrong phone numbers to all the wrong, uh,

people and, and then uploaded it, not realizing what I had done.

And then I realized that I had basically destroyed a.

A 2 million record Salesforce database.

And luckily, luckily I had a backup of that.

You are like, thank God for that.

W. Curtis Preston: So, so, well, what's the problem with that?

What's the problem with the status quo where we have all of the,

um, all of these copies around what, what's, what's the big deal?

There are multiple problems.

One is from a compliance governance perspective, you

need to know where your data is.

If this is production data, you need to know where those copies are, who has

access to it, who's spinning them up, how long are they being kept around for all of

these things, which become very difficult when you don't know who's creating

copies and where the copies are going.

So that's kind of one problem, just compliance aspect.

The other thing is you also want control over the lifecycle of those copies, right?

How easily can I spin up a copy?

Because if it takes me three weeks to copy over data, I.

I have another copy that I can now start to use for tests that's going to

impact my how quickly I can do testing and find issues and all the rest.

And so you want mechanisms that make it easy to create these copies.

And also along with the compliance piece, making sure

that you have proper retention.

You don't want these copies living around forever either.

Right, because this is a copy of your data.

It could be exfiltrated if someone, if you get hit by ransomware or a

attacker gets into your environment.

You don't want them pulling out like copies of your production data because

you didn't realize, hey, someone had squirreled away a copy over here.

W. Curtis Preston: Yeah.

And to go back to the earlier thing, you don't want 'em pulling out

copies of your development data.

If your development data is copy of your production data.

One of the, one of the reasons why you winced when I said that, uh,

of course one of the things that we talk about when we do pull.

Production data and put it into the development is this

idea of masking it, right?

So that you have product Yeah.

Sanitizing.

Yeah.

Uh, so that you have production like data, but not actual production data.

One other thing to add with copy data management is it's not

like you take one copy once and you're done forever with copy data management,

there's an ongoing lifecycle it.

Developers are constantly building new features, needing to test, right?

So it's not like you could take a copy of your production environment

from a year ago and continue to use it because things change.

And so you also need that ability to refresh these copies and make that

as automated and as easy as possible.

W. Curtis Preston: Yeah.

Agreed.

Um, the, the other thing I, I, I think you alluded to it in your.

You know, when I, when I was, when I was asking you about the, the, the

problems of the status quo, one of the things, at least when CDM vendors were.

Describing this, this new wonderful world of CDM was the cost of

all of those copies, right?

Did every one of those copies, I mean, you know, you talked about you definitely, you

want to, you wanna manage that process.

You wanna make sure that, that you, you know where all the copies are.

You want to make sure that people are, the, the right

people are accessing copies.

The wrong people are not accessing those copies.

Maybe some of the copies are encrypted, maybe some of them are not.

Um, maybe they have different performance, uh,

aspects.

Right, right.

Uh, yeah.

Performance characteristics, better word.

And the, but the, the other thing is that if, if all of your copies

are indeed on extra storage, right?

Or, you know, if, if each copy is on, on its own storage, that is

a very expensive process indeed.

Yeah.

And I know we talked the other day about snapshots and clones along with that, and

that's one way to sort of get your space optimization is by taking clones on the

same storage array that you have, the production or a copy of the production,

you can now quickly spin up those copies in a space efficient manner, right?

Rather than requiring hundreds of copies that are each occupying

the full amount of dataset space.

W. Curtis Preston: Yeah, so, all right, so that's our problem.

The problem is that we have way too many copies that we're paying

for way too many copies, each of which is its own attack vector.

And, and, and we still, we still want to protect all of that, right?

We want to protect it from a backup and recovery and

disaster recovery perspective.

Whoa.

W. Curtis Preston: how does cd, what, what did I

Just that last one about we wanna protect it.

I think it depends.

You may not

W. Curtis Preston: don't want to predict it.

well, the copies you may not care about your

W. Curtis Preston: Well, I just, I just, I just meant the data.

The data in general.

Yeah.

We don't necessarily want to protect each copy.

I just meant the data in general.

Um, you know, if we have nine copies of the different data, we

only want to protect one of them.

Right.

So,

W. Curtis Preston: So we're on, we're on the same

okay, good.

So.

Yeah, so just trying to understand the environment and I think this is

where it becomes difficult because differentPrasannaonas in your

organization use different tools and have different requirements.

W. Curtis Preston: Mm-Hmm

I look at, say your database admin.

They're gonna wanna spin up a copy, copy off a production in order

to be able to do testing, right?

Or to give to the database team for them to do application

development and all the rest of that.

And so they're gonna wanna copy off of that using their own tools.

So they might integrate with something that is more database

friendly, like Delphix, which allows for spinning up test and dev copies

off of Oracle and other databases.

Then you have, uh.

Folks who might wanna care about, okay, I need to be able to spin up

a copy to do like my backup team.

I wanna be able to spin up a copy to make sure that I can verify my backups.

And make sure that I could restore my data in case I need to.

So do backup verification.

And so in order to do that, I need to spin up a copy off of my backup

system in order to be able to access that copy, do my testing without

impacting the original backup.

'cause that's a key.

You don't wanna change any of the original data, you just wanna

copy of that data and manage it.

And so there's a separate lifecycle for that.

W. Curtis Preston: And I think that there are products like Veeam, and I

know Rubrik has done this quite a bit.

I believe, uh, Cohesity has done this quite a bit.

I, I know Druva did this as well.

Um, well, they, yeah, they specifically, Dr you know, since I

worked there, but specifically Dr.

For, for VMware.

But, uh, uh, so that's, you know, you talked about Delphix.

There are products that are specifically aiming at this.

Backup side of of CDM.

Yep.

And I think one of the biggest challenges you have is each Prasanna

is gonna use a different tool.

I know three, or like five, seven years ago, everyone was like, Hey,

we'll just have one single tool that can cover the entire environment,

that can manage copies everywhere.

And I think that's just a difficult problem to solve.

Right.

Uh, because you're never gonna have that one tool that everyone likes because every

Prasanna is going to want their own tools or gonna have their own custom workflows.

And so it's hard to say there is a single ring to rule them all.

W. Curtis Preston: Yeah, I, I think that's probably Actifio that, that

that's what Actifio was going for.

Right?

Um, they eventually got acquired, I forgot by whom I.

But, um, the, that, that was certainly what their goal was, right?

Was to, was to provide a copy, a copy for everybody.

It's like everything else in the IT world.

If, if you do everything, you're not gonna be any good at anything, right?

And so, and so, you're going to run into, you know, like the delphix of the

world or the, the beams of the world.

You're gonna run into somebody who's really good at that particular workflow.

And so, um, so, so it sounds like what we're saying is there's, at least right

now, we're not aware of any one tool that meets all of these copy data management,

uh, needs for every type of, uh, workflow.

that I'm aware of.

The only thing you could possibly do is have some sort of reporting tool

that's at least able to discover where all the copies are and kind of

stitch together a picture for you.

But that may not give you sort of the orchestration you need across everything.

W. Curtis Preston: We look at the world of copy data management, I, I

see sort of, I think there's like.

Three big things, right?

So one is the discovery, where is everything?

Where are all my copies?

And then the second one is who has access to those copies?

And then the third is, um, how long are those copies going to be around?

Because for, and I'd say that's primarily for cost reasons,

but it's also for risk reasons.

Right?

Because as long as something is sitting around, it's something that

can be, uh, potentially attacked.

Does, do you think that sums up the, the,

I, I.

W. Curtis Preston: we're worried about?

I know you talked about cost, but I wanted

to talk a little bit about that.

But yes, I agree with the three things that you summed up

W. Curtis Preston: Right.

Well, the cost is sort of the reason we care about the, the,

the three, the third thing, right?

That, that, that time.

Um,

Well, I wanna mention though, why copy data management, though

sort of has gotten better now and why it was talked about a lot more versus before.

W. Curtis Preston: okay.

One of the things, and also some of the

downsides of copy data management.

So one of the things is when you had sort of traditional disc spinning up

a copy was even if you had snapshots and clones, you could do that.

So you're not consuming extra space, but it would take a performance hit.

And when you started to look at SSDs, now you can actually afford to spin up

a copy and offer the same performance.

To those other application use cases without necessarily fully

impacting your production data source.

That's what a lot of people are concerned about, which is why they would copy

off the data to a secondary system and then spin up their test in Dev with

SSDs and high performance Flash arrays.

You no longer had to worry about that as much, and so you can now spin up copies.

Like if you talk to our friend Howard at Vast Data, right?

I'm sure VAs will say, yeah, just spin up as many copies as you

want off of the vast production.

You'll be fine.

So that's one way that sort of production copies became a lot more.

Affordable and also not as much concern from performance perspective.

Now, the other thing I will mention is backup target systems have always been

sort of like, Hey, we have all this data.

How do we add more value?

How do we allow people to use the data?

And I know we talked about backup verification testing or restore

verification testing, but people started then thinking about can we

start to use backup systems for test and dev scenarios and other things like

W. Curtis Preston: Mm-Hmm.

Now if you look at an all flash array system, which I don't

think many purpose build backup appliances exist, which are purely all flash just

because of cost reasons for those systems, they mix flash with a good amount of disc.

And so yes, you might be able to spin up 1, 2, 3 copies off of that system, but

at some point you're gonna run into a performance limitation much sooner than

you would off of your primary system.

The other thing also is.

People don't typically associate backup systems with doing these test

and dev copies because backup was built for a reason, and that's to

provide you sort of that last line of defense against everything else blowing

W. Curtis Preston: Right.

And so is that something from a risk perspective,

as a backup admin, you really wanna give the keys to everyone

who wants a test in dev copy to be running off of that backup system.

Or really is your main focus of that backup system?

Hey, I'm the last line of defense.

My goal is to make sure I can restore data, meet my SLAs, protect

data in the corporate environment, and test and dev really should be

run by a different organization or go talk to the production admins.

And so that's why I think copy dev or copy data management never really took off

as much as people wanted to from backup systems as it has from production systems.

W. Curtis Preston: Yeah, and the, and, and I think for other reasons, like

we want to keep you, you talked about.

Why not to use the backup system as a source for copies for test and dev.

And I'm also saying why not to think of the backup as just another copy.

Because when, when it's, when you know if we get that full CDM system, if backup

is just another copy, my question is how connected is it to this primary system?

That could go offline and, and does that in some way, uh, put my backup and Dr.

Copy at risk.

And my business at risk, right?

Because that is the purpose of the backup and Dr.

Copies.

W. Curtis Preston: Right.

And so I, I think that there's still a good argument to be

made for keeping backup and dr.

Copies separate and then production and dev copies perhaps together, uh, provided

off of the same storage because as, and as long as you do it in such a way

that you're not impacting production.

Yeah.

W. Curtis Preston: Um, if, if you're, and I, I think that with

flash, that does solve, uh, that or that ameliorates that concern.

Um, and then, you know, we haven't talked at all about cloud copies

Yes.

W. Curtis Preston: because I think that's where the true, like it's

so easy to just spin up another copy of something in the cloud.

Then forget about it, right?

It's so easy to make an S3 copy of something to make a, a copy

of something in EBS to clone a VM from here over to there.

Um, I think that when, when we, now, you know, early in this recording I

talked about, I used the word holistic.

You really have to bring, you know, when we say whole, the whole part

of holistic, we've been talking primarily about the data center, but

we really should be talking about those copies in the cloud because

it is so easy to create both backup, archive development copies, et cetera.

Entire instances, right?

Entire, um, virtual data centers That then no longer get used.

Yeah, and that's a big thing from like a cost perspective

because I'm sure there are companies out there which have a bunch of either.

Snapshots or EC2 instances running that they aren't even able to keep track

of because at some point you have so much infrastructure running, you don't

know what is being used for what.

And so that becomes a concern because that starts to eat

because it's not cheap right?

To run that.

It's not your own data center, you're co, you're worrying about, but you're still

paying AWS or Google or whoever your cloud provider is for those resources.

W. Curtis Preston: Right.

I think I'm agreeing with you that I don't think there is the one,

the one tool to rule them all.

Because there's a data center versus the cloud for no other reason other than that.

But there could be a tool that could manage, that could do

these, the three things, right?

Uh, find all my copies, uh, find out who has access to them and find out

how long they're gonna be around.

And then also I think in the cloud there are ways to find out if a particular

resource is actually still being used.

I am sure there is.

Yeah.

W. Curtis Preston: Yeah.

And so if we can, if we can identify these resources that aren't being

used and then get rid of them, and then I think that that's sort

of that initial, uh, challenge.

But then once you get that initial lay of the land, I think if you

have a tool that can then be the one who creates the copy for you,

Yeah.

W. Curtis Preston: right?

Uh, if it can create the copy for you rather than just

you as using native tools.

Um, then I think that you could get that cost aspect and risk aspect, I

think a little bit more in control.

and I think in the cloud also because of the ability to tag

resources, you get a lot more flexibility.

So you could say, Hey, I'm spinning up this test and

dev copy for this department.

You could tag it in your EC2 instance, and you could use that later to track

and understand, okay, who's using what?

So I think a cloud helps with a lot of the manageability aspects.

You just have to use it in the right way.

W. Curtis Preston: Yeah, I think tags are, I think tags are, um, you know, uh,

an amazing tool that, uh, are really, they're, they're helpful tags, are

very helpful for a number of things.

This is copy data management is one of them.

Another area where they're helpful is backup and recovery.

Um, because you can apply rules to different tags, right?

And you can say, we're gonna, we're gonna handle, um, our VMware cloud

instances like this, and we're gonna handle our, you know, our

EBS instances like this, et cetera.

And you can just, you, you can attach rules and backup and DR policies

to those rules, uh, based on tags.

And you, you can even have.

Um, you can say, Hey, if a resource gets created and it doesn't have

a tag at all, this is what we do.

The default.

W. Curtis Preston: One, one of which, yeah.

The, the default policy.

And, and, and part of that policy would be to go yell at somebody

to say, Hey, why is there, why is there this resource without tags?

Um, just to finish out, we, we talked a little bit about, uh, the reuse of data.

Um, I think that.

This, there are, you know, as long as I've been in backup, there's been

a lot of talk, especially in the last, I'd say 10 or 15 years, there's

been a lot of talk about trying to leverage backup for other purposes.

You've touched on it already.

Um, I, I think from a, you know, backup is different than archive.

I do think that if you stored the data correctly, if you stored

the data in such a way that.

You had information available for both backup and archive instances,

and you were able to meet both workloads with the same copy of data.

I would be fine with that.

I very rarely have found I.

You know, backup software.

I, I'd say about the only one that I've seen that does it, at least the best that

I have seen would be Commvault, right?

That they have a single copy and they have both backup and archive.

Now, the concern that I've, that I've heard is that it does

require a significant amount of additional infrastructure.

Um, but I, I, I don't have a problem with that additional cut because that I see as.

You're saving money by storing one copy and using it for

multiple

W. Curtis Preston: I just don't want people, right.

I, I just don't want people taking their backup and then holding it for

10 years and calling that an archive.

That is not an archive that we've, we've had episodes about that.

Um, I will talk about that till I'm blew in the face.

Um, and, and other, other things that we can use backup for is to.

Basically look at the backup and leverage it from a ransomware perspective.

We, we can look at and see if we can identify that the ransomware

attack has been, has happened.

Uh, we can see that if anything has been encrypted, we can see if anything

has been infected in the backup.

We can do that with the backup copy.

Um, but I think I agree with you.

Um, and, and potentially, and this is potentially, uh, if

we have backup an archive.

If we have that, potentially we could use backup for, uh, compliance

Yep.

W. Curtis Preston: right?

If, if we're able to easily query and, and, and this is a big if, right?

Because in order to meet compliance needs, you need to be able to,

the system has to act more like an archive system than a backup system.

I need to be able to say, show me all the files that Curtis made, right?

I need to say, show me all the emails with this word in them.

Show me all the documents with this word in

which most backup systems don't focus on.

W. Curtis Preston: which mode?

Backup systems.

Yeah.

Almost none, basically.

Um, there have been some advancements, uh, to, to, you know, to do some

stuff, but it's just not there.

Right.

And, and even then, it's often very.

Application centric.

Like we can do it for 365, but we can't do it for other stuff.

Right?

So again, if, if you want to use backup for archive purposes, I'm sorry, for,

for compliance purposes, then you, um, you just need to make sure that

it's capable of meeting that workload.

Yeah.

I agree.

W. Curtis Preston: Okay.

All right.

Well, we have once again beat this topic to death.

If you found this topic, uh, interesting and you wanna read more, you know,

similar topics, uh, you can get my book, modern Data Protection Available

wherever books are sold, including, you know, in Rivers in South America.

Um, some people don't like to buy books there.

That's fine.

I don't care.

I don't care where you buy my book.

Uh, all right, well thank you very much, Prasanna for a great discussion.

As always, Curtis, it was a lot of fun.

W. Curtis Preston: And thank you to our audience.

We'd be nothing without you.

And with that, that's a wrap.