Datos IO at AWS re:Invent
Transcript
[Music] hey how's the honesty bouncer from the CTO advisory joining you from beautiful Las Vegas aw sorry event P Peter it this has been a heck of a show so far fifty thousand people you guys have a booth data say an idol yeah and I stopped five briefly tongues with me to see huh I guess all you and I'm sure you didn't see me I thought I was gonna be able to stop by say hi but you guys were super engaged with folks coming by talk
about first you roll it diggle's I old okay and then the vibe of the conference cool so yeah peace male vice president of marketing business development for data sight oh we're cloud data management company and to your point about this is a great venue for us about 50% of our customers run in the cloud mmm but yeah I feel like I've been here I've been here since Sunday so it's been a couple days but I feel like I'm in here all week because it's just been
just the numbers are so big there's just people everywhere it's been a steady flow of traffic the buzz is good I didn't use this I didn't use this earlier today but it's like I feel like I'm back at contacts you know I mean I know that may not be a good thing because it's like context I have good at but but just for a magnitude it's sort of like the buzz and energy there's a ton right so I've been doing the cube interviews but and and
I've seen people flow back and forth matter of fact I had you on a cube earlier today that's great enjoy so this is a little bit different from which I talked to the CTO and Enterprise Architect Lea folks we had you guys on the podcast mother so the really great feedback from did a lot of a ton of downloads a lot of energy around that podcast yeah thanks for that that's great oh it was fun but since then you guys have come out with a new
version and then on top of that you're getting some insight last week didn't realize one you guys are not a cloud only solution right can you talk a little bit about the mix of cups the customer makes ya know what's new with the problem sure so I'll take a small step back because I kind of jumped past that before so we're we are a cloud data management company specifically what that means to us is we're really exactly two primary use cases one is backup recovery okay
and the second is dan mobility right so the genesis of our company and why we're doing this is because the world's moving to cloud we're all on board with what guys are on apologetically cloud absolutely we are steadfast in our view that the world's moving to cloud and the second piece is that the world is largely moving largely moving away from sort of the traditional relational data on pram model right database center view of the world so our focus is very much from cloud data management
specifically for non-relational databases and data sources and big data file system so our beds a lot of people oh you're back and recovery right absolutely we're backup recovery unabashedly but because backups on a new problem but it is for these new modern applications built on these new modern database platforms running into your point in hybrid or multi cloud environments so I think we need to tease that out because I don't think we've ever teased that out on video what it means to be a non relational
database okay and you know if you're a traditional infrastructure person you're thinking you know what backup is backed up Dana's what does it matter if it's relational not relational till you break that down for us a little bit what's the difference between the fundamental difference in challenges from backing up a relational database versus the now relational shorts there are a couple foundational things the difference between a relational non-relational database or a traditional database traditional relational databases were they're monolithic you put everything in one big stack
right they were what they're considered always consistent so at any given time you can basically say database tell me what you look like right now you know DRI was very prescriptive if I wanted to recover my dataset from 215 I turned an eye on my v-max arrange it I've been recovered then so the whole Warfel database at 2:15 right exactly because we inherent to the database was if you wrote a piece of data you didn't acknowledge that the data was written right until it was written
so everything was always consistent so but new non-relational databases they basically give up consistency for performance okay so they're designed by a definition to be far more scalable mm-hmm be far more performant to be more agile more flexible because you're essentially going from this monolithic database view of the world to more of a micro-services basically they'll out they're not gonna make me scale out the data they correct that creates that combination of consistency performance requirements performance at scale and sort of this notion of eventual consistency
creates underlying challenges for backup because you can't do simple things like a database tell me what you look like right now give me a picture that I have and then I'm gonna make a copy that I'll put it somewhere in I'm all set so look you know underneath his file systems I can just back up those file systems do some type of time sync and eventually consistent me myself what let us deals I'm thinking I'm guessing this how do you guys kind of make that simple
or what where's the fallacy in that in that concept the fallacy is in the cost is in the complexity and the risk associated with doing that I mean sense could you the could you come up with the way like there are there are there are customers out there today ideally some of them listen to this you know listen to the DOS and basically it's like you know are you running any date you know Cassandra databases yesterday you're running in the production yes we are what are
you doing for backup recovery we just run scripts right or we just do you know we just basically take snapshots of all our nose and then I mean when we Steve stacks together when it's time if I need to recover what's at and what will happen typically in those snares one is that's a very complex process it's a very expensive process and then a recovery standpoint you may have to go through a very manual time-intensive recovery you know recovery process repair process excuse me to essentially
rebuild that consistency contrast that with what we're doing which is completely automated fully orchestrated any point in time backup recovery you know always consist in any point in time recovery so essentially it's its point in time it's it's completely automated it's simple point-and-click and it's incredibly storage efficient as well because as we talked about in previous sessions traditional deduplication which everyone who knows backup recovery is familiar with the application that's a block level construct yes that does not work with you know with encrypted data with
objects story right with with eventually consistent or excuse me with clustered data sets it just doesn't wait for was an architect of theft so we're and we invented something called semantic deduplication okay which is specifically designed to deal with compressed and encrypted data I sure didn't cluster data so and how do you distribute databases that can have same data somewhere else if I'm backing it up with a traditional solution I'm going to deep duping that it becomes a bigger challenge you guys naturally do that great
example oh I can I can double click on that very briefly which is you know non-relational databases are clustered right but the data that's replicated across those clusters isn't identical from a block standpoint so just simple examples but I've got three data but when I bought but not blocked it's not blocked consistent and it's potentially encrypted or compressed just kind of throws traditional didi about the window because we integrated the database level and we hope semantic deduplication on purpose we understand the underlying schema of the
databases that that you're storing your data in so we can do key value to duplication so this gets us back to the mix of customers part of my original question there's one premises opportunities and there's off prim there's stuff that's in the cloud yep I can easily conceptualize this and running in a cloud Cassandra database etc what if I have those things on print how many you were customers are taking advantage of that technology cramming what does that look like from a packaging it looks identical
I mean not literally identical but essentially it's the only the layer that we care about is the day data source so because we integrated the data store at the database level we're actually agnostic to the infrastructure even we don't look at London so we don't look at the enemies we don't look at containers we literally integrate directly at the database level so as I mentioned about 50% of our customers run in the clap across multiple different cloud platforms the other 50% are running these modern as
we call modern applications built on these non-relational databases they're running on proud and so we'll run on bare metal we'll run a VM you know just as easily as we'll run in an ec2 instance so it's actually transparent to us the typical scenario for an on-prem environment is they'll either run it off bare metal or sped up the VM and then the storage they'll typically use is either NFS cuz will support any NFS or object storage so depending upon what they want to do they can
just use off-the-shelf nest right or they can use object storage and then in that same scenario then in the clouds AWS will be spun up as an ec2 instances an ami they'll use they'll be running through their clusters and then they'll use s3 for storage so let's close out these doses I really enjoy having folks like datacite old sponsor of these events really helps community learn more about the show and what's going on in the industry list close out one of the show what are you
guys most excited about it this AWS there's 18 billion dollars that's metric but 18 billion dollars was before you sink a year-over-year growth that's insane you look at kind of other vendors how you know talked about HPE HPE is shrinking there are a 30 billion dollar market cap company now if AWS was a separate company they'd be what uh easily company the scale is amazing almost 50,000 people at this show where is I think the question people had to ask where is the white space when
you look at something as massive as a AWS entering a market I'm sure they'll into data protection at some point where's the value that you if you guys talk to customers that just can't be replaced you guys talk about the four-year head start yeah where where's the key value wasting a white space for us is backup recovery and data mobility for non-relational database applications and the mobility is a big part is us AWS isn't gonna come into the market and say hey let me help you
move your data to Azure right no exactly and they're always you know we're not on migration so no not angry always like you know data bus that's all right down no we're not gonna help you spin up a copy of your data in another cloud provider to run tests they'll fall off but what excites us most about AWS ones they're great partner right one number two is we're very strategically aligned because they recognize the value what we're doing they recognize the space that well they've got
the Rd outside your cover this whole non-relational did in AWS is DNA there are non relational gray you know so they see the proliferation of updated I mean the number one number two database in the cloud on AWS is MongoDB number three databases Cassandra so it's all about non relations so the white space and the value the value that we offer to customers is enterprise backup and recovery for these modern applications running on these modern databases the value in white space that we're covering for AWS
is we're delivering that enterprise backup recovery capability for those customers those enterprise customers that want to be hosting these apps your green ice it harder value to your partner in your customers that's what excites us most is we've got a very strong synergy with AWS in this you know and the other piece of the puzzle is if you ask the AWS folks what's really exciting is that backup recovery is a core requirement because it's such a core tenet of what enterprises do is that they get
it they're interested that's what we do nice energy so Peter I really appreciate you joining the CTO visor doles at AWS Reid event where can people find out more about details and data file and you worry if people are yeah so if you're at the show if you check this out of the show we're in booth 20 and 25 in the main hall it's been jumping so come find us or check us out at www.sailrite.com you can find me on the web at CTO visor twitter
the CTO advisor comm talk to you next CTO bills
