00:01
Mm Thank you everyone for tuning into this session. Uh My name is Vinayak and I'm the senior product manager for replication slash data protection at pure storage. Today we are going to cover a topic active D. R. Which is the included capability for providing disaster of resolution with flattery and we have a special guest with
00:33
us today, kim thanks Lord for being on this session with me and thanks a lot for talking about DR B. R. And how your organization is using actor. Br Hi vin, thank you for having me. Alright, so in dumps off, getting to know a little bit about your background. Why don't you share with our viewers a little bit around yourself,
00:57
your organization and what does your team do in that organization? Sure. I'm Kimberley legroom. The Chief information officer of the City of New Orleans. Our organization uh is the information technology and Innovation department and we service about 4500 city employees. We support not only the administrative part of
01:22
government but we also support our public safety teams. So police fire, emergency medical services um are all part of the support that we provide. And we deliver stable technology services throughout the city. So whether it is infrastructure and communications or data, data recovery, web and video services, uh customer support helpdesk.
01:43
We kind of serve the city's technology. So we are the city's Internal technology department. That's wonderful. So I would think you provide services for the most critical of the workforces for the city of New Orleans. From the recreation parks and the canoe service all the way up to the public safety teams. We respond to events,
02:10
disasters um with the best of them but our public safety teams, especially our police and fire units do depend on us for technology services. So mission critical at times, sometimes it's just routine services but we keep the lights on here. I see, I see. Alright, that seems very very mission critical.
02:35
So in terms of the youth case because we're talking about active er it refers to D. R. Use case. I want to ask a little bit about your requirements on your use case of how you intend to implement any kind of disaster recovery data protection solution in your organization.
02:59
I'll give you a little bit of background about our organization first. Uh and that kind of drives how we have to provide or what disaster recovery services look like for the city of New Orleans. We are a coastal city, we experience hurricanes on a pretty yearly, almost yearly and very regular basis. Um We host major events here all the time.
03:25
So we are a large tourist destination and post large events like Mardi Gras. Um And the Mayor's previous term in the last four years we've had more emergency declarations than in any other that any other mayor has ever had. Uh we had at one time 19 open emergency declarations. So that means scenarios present themselves to us very quickly and they could be of any kind
03:53
of any nature. So um disaster recovery is really critical to uh the way that we operate because um we are often recovering from one event type or another. Um just the nature our geographic location and the nature of of of where we are and then the climate that we're in we have to be able to move quickly and there are always threats to our physical environment that we're working
04:20
against. So D R. For us has to always happen. It's an evolving process and we just need to always be in a position to be responsive. That that's that's what makes art our disaster recovery strategy a very iterative, very iterative.
04:43
That's very important to understand, I think that your geographic location itself demands that you be just ready for D. R. Solution and for D. R. Recovery. I think our viewers were the it's important for them to understand that you know, there are geographic locations who face these kind of challenges just because of where
05:09
they are situated and located and they have to be quite robust in selecting our d our strategy, just like you on top of it, you are providing services for such mission critical teens and the departments that you have to be very very careful in selecting the strategy. So let's dig a little bit deeper into that in terms of what kind of
05:36
applications from a technology perspective kind of depends on your d our environment. What kind of applications are the ones that you try to protect in such disasters? Sure everything voice. So very time sensitive applications. We support, our communications infrastructure is supported there um connectivity to other cloud based critical applications we also provide but we store great
06:07
amount of information for the city of New Orleans and we host all of their external services. So we're doing everything from hosting, external websites delivering emergency notifications or keeping communications channels open. Um As as with other organizations Much uh several of our key applications have moved to the cloud so we do have other parts of those
06:33
applications hosted somewhere else but our communication to and our access to those those locations are to those systems is is really critical for us. We have about 130 sites in our network. So in this in this area are metro ethernet network is pretty diverse. Um we centralize those services. So the timeliness of getting services to all of those locations is a service in itself.
07:01
And then the applications that we run over that are pretty critical either to public safety or when it comes to having immediate access to information. I think I think we are handling pretty large amounts of information quickly. I see I see. And then what is your D. R. Procedure today? So D. R.
07:22
Procedure today? Um last year we had a very different D. R. Procedure. We we had a different infrastructure but we were affected by hurricane IDa last year and we had a fire in our data center. Um while we were responding to a physical threat, a weather related threat we had an internal threat. So that meant that our d our strategy changed
07:48
because our data center location had to change immediately. So now our strategy has moved from an on site data center and a geographically dispersed data center two to geographically dispersed data centers and primary data center being a remote site. So our strategy now is um remote to remote um and having something very active in between those two sites is really critical for us.
08:16
And also having good communication between our physical locations to our data centers is really important. So I think what you are mentioning is earlier, it used to be same data center which had like your primary and then D. R. Backup as well. But then due to the experience that you had last year you decided that this should be
08:39
geographically apart now and then you kind of split them. Okay well we didn't even decide we were forced into uh into a situation where we had to move into our remote data center and make it our primary data. And how far are you do this interest today Right now they are about 60 miles outside of the city and the other one is about 300 miles away.
09:05
I see. All right so that way you'll be able to at least protect against fire kind of situation and hurricane kind of situation? Both? All right now in terms of your D. R. In terms of your R. P. O. R. T. Requirements because those are kind of critical in terms of disaster recovery, what do they look like for your organization?
09:29
Um So we've always been real realistic. We are a government. Uh So and and government um active Active is about cost. Right? And an organization of our size in the city of of our size, we've never been able to deliver a within hours R. R. O. I. R. R. P. O. So we we have been realistic with our
09:55
leadership about about what that looks like. Um And I would say that in the past We've gone from a seven day scenario down to about 48 hours scenario um when it comes to recovering data. Um many times we have so many threats or other dependencies on our recovery points and our our recovery time objectives that We we have to stay within. What we think is realistic for the amount of
10:24
infrastructure that we have. So right now we're at about a 48 hour recovery. So kim I think our viewers would like to know a little bit about how after B. R. Is used in your organization and what you have found it as a beneficial solution compared to others. Well, so with our with our multiple data centers now and our primary our remote data
10:49
center becoming our primary data center, we've had two geographically disperse our data are secondary data center somewhere else. So with active D are we have to flash arrays uh and in those data centers and what what has been great is being able to stream uh streaming nature of the active D. Are going from one storage cluster to the other storage cluster. And I know there is some tolerance for network
11:16
delays, but it really keeps our data in sync in an almost uh and in a very active way. So um Even with the disbursement of of our sites, they are about 300 miles apart and the tolerance for uh working over that network in any latency, it produces the fact that we're constantly streaming. That data means that that data stays in in sync in a way that just makes our recovery,
11:45
makes us really comfortable with being able to replicate our data or have that data that data in our secondary data center available to us. That's wonderful because it's true that Active D. R. Is a continuous replication solution. So you get that continuous stream of data being shipped from your prior to D. R.
12:06
And because of that, your R P N. R. T. O. Usually would be uh not R. T. O. But our P. O. Would be near zero as well. So that gives that advantage. And we and we see less interruptions backups usually can be disruptive to other parts of our process and an organization like ours, It's 24/7 um the constant nature of the
12:30
replication or the streaming really means that that is continuously happening. Yes, and I think I want to touch a little bit about the network tolerance aspect that you mentioned too because that's kind of built into the product which essentially was baked into the product in case that if there is a network latency increases, like if there's some network that is lower or some disruption,
12:56
the product can automatically move into a different replication mode to catch up and then start continuously streaming as well. Yes, and that, that is one of the biggest factors in moving data centers and for us it had been one of the things that we that prohibited us from really implementing a good solution before network dependencies and latency um really has to be factored into a,
13:24
a solution and many solutions don't tolerate it very well. So this, this really worked well for us. That's wonderful. Alright, so besides actively are, did you also consider some other solutions while we were considering just the from our flash arrays, we considered other backup solutions,
13:46
Yes, we did. And in those, in those backup solutions we just found them to be more compartmentalized. Right and it meant that we had 11 part which was our storage and then we were very separate and we had come from an environment where our backup solutions had very separate dependencies are very operated very differently than the storage itself. So yes, we we decided to look at something more
14:14
integrated and pure storage offered us active D. R. As a part of our of the solution and it just made so much more sense okay now because active er provides d our testing capability, I wanted to touch upon that a little bit that do you do disaster recovery testing on a in your D. R. Strategy as well or do
14:45
so twice a year? Uh And before the beginning of a hurricane season which starts in New Orleans May 1st we consider active hurricane seasons beginning in july also live first. So between May and june of every year we are testing and readying our d our solutions, we perform a similar test post D. R post hurricane season just to make sure that that we are aligned and many times just like
15:14
last year we were in a position where we needed to re architect or change some parts of our solution. So the during after the disaster is usually a good time for us to test any changes that we've put in place. I see. So in that kind of scenario um the workflow of activity are where you can do the our testing without disrupting your production workload
15:38
would have been very useful. Absolute. And that that is one of the biggest advantage is that we can test without interrupting our day to day services. We are 24/7 shop because of our customers. So while we make a home at night um our police and fire departments and our medical services, they are actively working. So having a scenario where we can test actively and throughout our day without interrupting
16:07
those services was really important. Garden Garden. Alright. So uh let's jump into some business value and business advantages that you guys have realized with pure and active er usage would you want to share a little bit about, you know what kind of time to value? What kind of like you said non disruptive we are testing,
16:32
how does that translate into business value? So for us the business value is continuity of services. Um That has been the most important part of our work. Um I would I would have to say no. Um Apart that our team really appreciates is the simplicity of the solution. So complex solutions require more resources for
16:56
us, they require more training for our staff. So being able to implement a solution quickly uh and have staff to understand it easily means that they can be more effective at their work. So what we find is we have less people that we need to have involved in D. R. They can focus on more critical or other parts of our data center applications. Um And because that solution has been easy to
17:24
learn, we've managed some transition and even some turnover but found that the solution has been so consistent that it's really been a resource saver for us here. I think along with just the way it works. I think the resource um the savings and the efficiencies and our resources because of the simplicity of the solution has been really um a really good investment for the city of New
17:53
Orleans as I mentioned. Um we're government so um that allowed us to make the investment in the flash array uh and really focus on understanding what type of data we needed to secure and having a simple activity, our solution like we have made our recovery and the people who actually performed that recovery, it made that much, much more affordable for us and it just proved its value to us as an
18:22
organization. I see. So would it be fair to say that time to value because of simplicity, lesser resources in terms of training and expertise and other than that the business continent that it provides. Those three have been kind of the key business values that you have realized by using flash array and actively are Yes,
18:47
yes. I would say that there could not have been a better scenario for us to have really tested this solution but to have two engineers and two days to have an environment moved and stood up and actively working was a great value to the city and it was something that we needed to happen in in real time and I can say we really didn't miss a beat.
19:14
Uh the storage and the solution worked as it was demonstrated to work. D. R. Scenarios and tabletop exercises provide some level of comfort to you. But we actually walked through this and founded to do to perform as we expected. That's great and and like you said it works and it works and that's why
19:40
it's the right solution. So we are at the conclusion of our session, would you mind sharing a little bit about contrasting it with how the D. R. Solution or process used to be and how it is now with flattery that I think that would give our audience a little bit in terms of what it used to be and what it is now and how it's different.
20:09
So 2019 before we had pure storage uh I'll tell you in very, very real terms are 70 terabytes of data that we recovered took us months. We actually broke it down into those categories and tears and we prioritized that data. So we went from a a multiple hours of multiple day scenario for each of those categories of data which took us about about
20:38
three months to identify clean and store those those um that data but as I mentioned the storage was very different from the from the application. So we experienced a lot of complexities and making those applications work Uh fast forward to 20, same amount of data, maybe even more, I want to say about 100 terabytes of data that we were recovering. We went from months to 48 hours Uh and it was
21:09
really a matter of bringing online data about 100 terabytes of data and about 300 virtual server environment. Um and I think we did that in a fraction of the time compared to the solution that we had before. Wonderful thanks kim is there anything else any words of wisdom that you want to share with our audience?
21:33
Well, I I think you can never be too prepared but you know, the support that we got through the engineering team, the pure storage team, the demonstrations, um I don't think there was any more comfort that our team could have had, that what they saw and what they expected to happen actually did happen. So being prepared is really great but really understanding those solutions and the ones that
21:59
you employ and how they work is just so important to having a successful disaster recovery strategy. Whatever it might be wonderful. Thanks Lord kim for joining us today and for sharing your experience with your shortage and with active D R. As a D. R solution. I am immensely thankful for your time today and
22:22
I'm pretty sure that our viewers will find this very, very useful. They always love to hear from an actual customer how our technology is being used and how it provides value and benefit to our customers. Thanks a lot. Thank you, vin. It was a great conversation.
