# Solution Architect Technical Interview (Master the Solutions Architect Interview Questions)

https://www.youtube.com/watch?v=8IJUHf5cdbc

[00:00] are you looking for solution architect
[00:02] interview questions or cloud architect
[00:04] interview questions if so this video is
[00:06] for you
[00:07] [Music]
[00:13] hi my name is michael gibbs and i'm the
[00:15] founder and ceo of go cloud architects
[00:18] and we're an organization that's
[00:20] dedicated towards building high
[00:22] performance cloud computing careers
[00:24] personally i've been working in
[00:25] technology for over 25 years and i've
[00:28] been coaching others or mentoring others
[00:30] to get their first tech job or get
[00:32] promoted in tech for over two decades
[00:35] today we're going to talk about
[00:36] preparing for that cloud architect
[00:39] interview or that solutions architect
[00:41] interview and we're going to do so by
[00:43] giving you some solutions architect
[00:44] interview questions or cloud architect
[00:48] now
[00:49] when you're asked these interview
[00:50] questions from the hiring manager
[00:52] they're asking it for a few reasons
[00:55] first and obviously they're trying to
[00:57] check your technical competency but they
[00:58] could do that in a lot of different ways
[01:01] the reason they're going to ask you a
[01:02] lot of open-ended questions they want to
[01:04] not only see your technical competency
[01:06] but your ability to describe technology
[01:09] your communication skills as well as to
[01:12] see how you act under pressure so what's
[01:14] the best way to act under pressure be
[01:16] prepared for the pressure in the first
[01:18] place if you're prepared it'll be
[01:21] a review when you're there not the first
[01:22] time so you won't be stressed so that's
[01:24] why we have these videos on cloud
[01:26] architect interview questions to make
[01:27] sure you're successful on that cloud
[01:29] architect interview or that solutions
[01:31] architecture
[01:32] the first question that we're going to
[01:34] ask you is to describe
[01:37] the four types of disaster recovery
[01:39] options like cloud and the strength and
[01:42] the weaknesses of each approach
[01:44] so first
[01:45] obviously you need to know the four
[01:47] types of disaster recovery options on
[01:49] the cloud and then you need to know the
[01:50] reasons why an organization would choose
[01:52] each so let's walk through that together
[01:55] the first
[01:57] disaster recovery option is just simply
[01:59] backup
[02:00] let's say you're on a cloud provider or
[02:02] your data center you take an image of
[02:04] your servers and you move it over to the
[02:06] cloud
[02:07] you back up your data and you back it up
[02:10] to the cloud and maybe you say once per
[02:11] day you take your data from the cloud
[02:13] and you sync i mean your data center and
[02:15] you synchronize it with the cloud that's
[02:17] it that's back up
[02:19] what's great about this you back your
[02:21] data up to the cloud
[02:23] it's very cheap and if you needed to get
[02:26] the organization's systems up and
[02:28] running
[02:29] within about 12 hours you could do it
[02:31] the organization is completely up it's
[02:33] been a wonderful low cost
[02:36] high environment way to do
[02:39] disaster recovery high reliability and
[02:41] extremely low cost okay so let's talk
[02:44] about the next time
[02:46] in the previous version
[02:48] we talked about copying your data and
[02:49] synchronizing can say every 24 hours
[02:52] so the next version would be to keep our
[02:54] data synchronized a little more
[02:56] frequently but not all data needs to be
[02:58] synchronized a little more frequently so
[03:00] let's talk about how we can do this
[03:03] in disaster recovery version 2 we're
[03:06] going to take the same images of our
[03:08] servers as we did previously and move
[03:10] them over to the cloud we're going to
[03:12] back up our data say every 24 hours and
[03:14] move that to the cloud but in this case
[03:16] we're going to keep a database
[03:18] synchronized between say our data center
[03:20] and the cloud and by keeping our
[03:22] databases synchronized between the data
[03:24] center and the cloud guess what now our
[03:26] transactions are synchronized in both
[03:28] cases so if there's an outage in our
[03:31] main environment say our main data
[03:33] center we go up and running on the cloud
[03:34] it's still going to take us about 12
[03:36] hours to come up but our transactions
[03:38] are synchronized so our data is fresher
[03:40] so it's an improvement over disaster
[03:42] recovery option one
[03:45] the next form of disaster recovery
[03:47] disaster recovery option three the cloud
[03:49] really shines here and this is because
[03:52] the cloud has auto scaling so disaster
[03:54] recovery option three really revolves
[03:56] around auto scaling and what this method
[03:59] is is in your data center you've got
[04:02] your actual real environment and you
[04:04] create a replica environment in the
[04:06] cloud meaning if you had a network load
[04:08] balancer and a hundred web servers in
[04:10] the data center maybe you set up a
[04:12] network load balancer or two and two web
[04:14] servers in the cloud and is in an auto
[04:17] scaling group so that can scale out as
[04:20] needed and say you do this for all of
[04:22] your environments then here's what
[04:24] happens
[04:25] if anything happens in your data center
[04:28] traffic will be redirected to the cloud
[04:30] and all your systems are running in the
[04:32] cloud and they're all synchronized but
[04:33] they're in auto scaling groups and
[04:34] they're small so all the traffic would
[04:36] be redirected to the cloud the cloud
[04:38] would auto scale systems would scale out
[04:41] in about 45 minutes to an hour the cloud
[04:44] would be a perfectly operating fully
[04:46] fully established disaster recovery
[04:48] environment for the organization so
[04:50] disaster recovery 3 is really amazing it
[04:53] provides high speed high performance
[04:55] failover for disaster recovery and it
[04:56] does it at a relatively good cost by
[04:59] leveraging small instances of everything
[05:01] on the cloud a replica environment and
[05:02] using auto scaling
[05:04] now the last form of disaster recovery
[05:07] is the pure pure pure
[05:10] active active
[05:12] and here's what this does if you've got
[05:14] an organization that's got a thousand
[05:15] web servers in their data center you're
[05:17] gonna have a thousand web servers in the
[05:19] cloud if you have a thousand app servers
[05:21] in the data center you can have the same
[05:22] thousand app servers in the cloud it's
[05:24] just gonna be a mirror image and if
[05:26] anything goes wrong in the data center
[05:28] the traffic will just get redirected to
[05:29] the cloud and everything's running so
[05:32] what have we got we've got backup we've
[05:34] got backup plus synchronizing our
[05:35] databases we've got a small instance
[05:37] version on the cloud using autoscaling
[05:39] and we've got a complete active active
[05:41] hot hot environment those are your four
[05:44] types of disaster recovery options on
[05:46] the cloud
[05:47] the next question we're going to ask is
[05:49] a networking question and it's a really
[05:52] really important question
[05:54] most of what we cloud architects do is
[05:56] we take systems from the network and the
[05:58] data center and we migrate them over to
[05:59] the cloud
[06:00] now what are we migrating we're
[06:02] migrating data
[06:04] as well as systems
[06:06] and the way we get those data or systems
[06:09] to the cloud could be one of a few ways
[06:12] if we've got a private line or a direct
[06:14] connection or a vpn we can just transfer
[06:15] it over
[06:17] but if we don't have enough capacity on
[06:18] those lines or we have too much data to
[06:20] transport and not enough time we might
[06:22] need services such as the import expert
[06:24] service the snowmobile or the snowball
[06:27] and the capacity and how much data we're
[06:29] going to have to transfer is going to be
[06:30] there but you're never going to know
[06:32] this if you don't know the capacity of
[06:34] your link so the next question is
[06:37] how much data can you transfer in 24
[06:39] hours on a gigabit ethernet link
[06:43] so i'll say it again how much data can
[06:45] you transfer in 24 hours on a gigabit
[06:51] so
[06:52] you will see this on exams like the
[06:53] certified solution architect
[06:55] professional but you will see this on
[06:56] interviews so how do we determine this
[07:00] well we know that a gigabit per second
[07:03] is a thousand megabit per second
[07:06] so we also know
[07:08] that there are eight bits to a byte
[07:11] so let's do this first
[07:13] if we take a thousand megabits per
[07:15] second because it's mega bits we first
[07:18] convert that into bytes so we take a
[07:20] thousand megabits divide by eight and
[07:22] now we know we can transfer a 125
[07:25] megabytes per second
[07:27] well there are 60 seconds in a minute so
[07:29] if we take our 125 megabytes
[07:32] and we multiply times 60 we know that we
[07:36] can now transfer on this link 7.5
[07:39] gigabytes
[07:40] per minute
[07:41] now there's 60 minutes in an hour so if
[07:44] we take seven and a half gigabytes which
[07:47] is the amount of data we can transfer in
[07:48] a minute times 60 minutes that equates
[07:52] to 450 gigabytes per hour
[07:56] now there's 24 hours in a day so we take
[07:58] our 450 gigabytes multiply times 24 and
[08:02] we get 10.8 terabytes so we know at
[08:05] least theoretically we can transfer
[08:08] 10.8 terabytes on a gigabit ethernet
[08:11] link
[08:13] next question
[08:14] why can't we get the full 10.8 terabits
[08:17] on a gigabit ethernet link
[08:20] now this is where we're asking your
[08:22] networking knowledge so
[08:24] what we're looking for here is we want
[08:26] to know that you know about the way the
[08:28] network works
[08:29] meaning
[08:30] just because we have
[08:32] a gigabit doesn't mean we can get it so
[08:35] for example if there's tcp and there's
[08:37] flow control going back and forth and
[08:39] acknowledgments that will cost us
[08:40] bandwidth but when we take the packets
[08:43] or the when we take the data and we put
[08:45] an ethernet frame on there the ethernet
[08:47] frame as a header adds overhead then we
[08:50] slap an ip header on there for the tcp
[08:52] networking and that has overhead
[08:54] we might actually have a vlan tag or
[08:56] something like that and that also adds
[08:58] overhead so the reason we will never get
[09:01] maximum performance on our link is due
[09:03] to the overhead and the encapsulation
[09:06] methods of ethernet ip in other words so
[09:08] there's always going to be about five to
[09:10] ten percent overhead on the link so now
[09:12] you know why you can't get the full
[09:14] utilization out of a link because
[09:15] there's an overhead on the way
[09:18] it's the same reason that if you have an
[09:20] 80 terabyte hard drive and you format it
[09:22] you might only get say 72 percent 72
[09:24] gigs in capacity like an aws snowball
[09:26] because the overhead associated with it
[09:30] in this next question we're going to
[09:32] talk about three kinds of storage when
[09:35] do you use object storage when you use
[09:37] block storage and when do you use file
[09:39] storage
[09:41] first let's talk about object storage
[09:45] you should be able to tell the employer
[09:46] that object storage is the type of
[09:48] storage area network that's very unique
[09:50] in the way that it operates and then it
[09:52] takes data and breaks it down into
[09:53] objects and that each object has
[09:56] metadata or data about that object and
[09:58] because each object has metadata about
[10:00] that object it is very easy to search
[10:03] very easy to query and can integrate
[10:06] really well into a big data environment
[10:08] you should be also be able to tell the
[10:10] employer that object storage is not
[10:12] computer systems like regular storage
[10:14] and that it does not get used by
[10:20] not suitable for regular computer
[10:22] systems because anytime something is
[10:23] modified even a little bit it would
[10:25] create a new version and that's why
[10:27] object storage could not be used as like
[10:29] a hard drive whether it be a swap file
[10:31] or an operating system or anything that
[10:32] would constantly be changing you might
[10:34] also want to be able to say that object
[10:36] storage is really more like a database
[10:39] and that it's not hierarchical in nature
[10:41] and basically the data is just placed
[10:43] into object storage and there's like a
[10:45] database pointer that points to the
[10:46] location of the object storage you might
[10:48] even want to be able to tell them that
[10:50] object storage is used for software
[10:52] distribution it's great for backup and
[10:54] archival purposes and it's great for
[10:56] data lakes
[10:57] so that's what you realistically want to
[10:59] tell them about object storage you could
[11:01] also call it s3 if you're dealing with
[11:03] awf
[11:04] you could also call it cloud storage if
[11:06] you're on the google platform that day
[11:08] and you could also call it blob if
[11:10] you're on microsoft but realistically
[11:11] speaking what we're just talking about
[11:13] is object storage the next type of
[11:16] storage
[11:17] is block storage
[11:18] now block storage is another type of
[11:20] storage area network technology where
[11:22] data is broken down into blocks now what
[11:25] makes block storage so good for the
[11:27] cloud providers is it effectively
[11:29] enables you to place the storage
[11:31] environment or the blocks
[11:33] anywhere it needs to be in the storage
[11:35] environment so by doing so it
[11:37] effectively decouples your compute from
[11:40] your storage so that's why your cloud
[11:41] providers are using block storage now
[11:43] block storage is network storage so it's
[11:45] not going to be as fast as local storage
[11:48] and its speed is going to be limited and
[11:49] its throughput is going to be limited by
[11:52] the network
[11:53] the type of network you're using so if
[11:55] you're limited at one gig or 10 gig or
[11:57] 100 gig that's going to be the
[11:59] limitation of performance that you're
[12:00] going to get from your black storage
[12:01] because it's network storage now also be
[12:04] able to describe the block storage
[12:06] looks and feels when it's mounted just
[12:08] like a hard drive
[12:10] so block storage is used in the cloud
[12:12] computing environment when organization
[12:14] needs something that would function as a
[12:16] virtual hard drive why did the servers
[12:18] in the cloud need a virtual hard drive
[12:20] because the servers come with basically
[12:22] the storage as part of it it could be
[12:24] called instant storage it could be
[12:25] called ephemeral storage basically what
[12:27] that means is the storage that comes in
[12:29] your virtual machines which is very fast
[12:31] goes away with system reboot so if
[12:34] you're going to have a server in the
[12:35] cloud and the server is going to need to
[12:36] have anything that's stored on it you
[12:38] have no choice you're going to use block
[12:40] storage because you can't store it on
[12:42] the instance itself so that's why
[12:44] organizations use block storage and
[12:46] that's why cloud providers use block
[12:48] storage because it scales so well the
[12:50] next type of storage is really network
[12:53] file storage and if we there's really
[12:55] two kinds of network file storage if
[12:57] we're dealing with unix and linux
[12:59] systems we're really dealing with some
[13:01] version of the network file system that
[13:03] was invented by sun micro systems now
[13:05] oracle you know a while back or we're
[13:08] dealing with some form of server message
[13:10] block for windows system so
[13:13] if we're dealing with aws for example
[13:15] we're going to have two options we're
[13:18] going to deal with their version of nfs
[13:20] called the elastic file system and we're
[13:22] going to use the elastic file system
[13:24] when we've got a bunch of linux and unix
[13:26] servers that need to look at the same
[13:28] information they're all going to mount a
[13:30] shared drive and that's why we're going
[13:32] to do that for our linux and unix
[13:33] systems so shared information used by
[13:35] lots of servers we're going to use efs
[13:38] and the aws cloud nfs and the data
[13:40] center
[13:41] now
[13:42] let's say we've got a lot of windows
[13:44] systems
[13:45] well
[13:46] we could obviously set up a server and
[13:48] run samba on it but if we're in the aws
[13:50] cloud they've got fsx which is windows
[13:53] servers that are basically a fully
[13:54] managed file system for windows so
[13:56] basically windows file servers so those
[13:58] are the storage options you have in the
[14:00] aws cloud and the purposes for each one
[14:03] now this next question is going to be on
[14:05] dns
[14:07] and we're going to base it on the aws
[14:09] flavor of dns today and we're going to
[14:11] say
[14:12] describe aws route 53 and the main
[14:15] routing policies and what they do
[14:18] first you need to know that aws's brand
[14:21] or the amazon brand of dns is called
[14:23] route 53 interestingly enough
[14:26] dns is tcp and udb port 53.
[14:30] so you also should know that dns
[14:33] basically maps a name to an ip address
[14:36] in its overall function basically giving
[14:38] you a name like
[14:40] www.go cloudcareers which is really easy
[14:43] to remember as opposed to its ip address
[14:45] so now let's talk about the types of
[14:47] routing so the first type of routing is
[14:49] simple routing this is the first type of
[14:51] simple routing policy what this is is we
[14:54] map an address
[14:56] to a name www.gocloudcareers.com
[15:01] that to whatever it's ip addresses
[15:03] simple routing the next type of routing
[15:06] policy with route 53
[15:08] is the failover routing policy this is
[15:11] quite simple
[15:12] we've got two data centers
[15:14] there are two servers somewhere
[15:16] send your data here
[15:18] if this goes away
[15:21] send the data here
[15:23] so just go to the primary the primary
[15:25] business go to the secondary that is
[15:27] failover routing policy here's what
[15:29] happens the system send a health check
[15:31] basically dnx is here and it says a
[15:33] message hey are you there are you there
[15:37] and the data center says i'm here i'm
[15:39] here i'm here and as the dns keeps
[15:41] sending a health check saying are you
[15:42] there the data says i'm here if for
[15:45] example the data center stops responding
[15:47] to health checks are you there are you
[15:48] there are you there no response data is
[15:51] shifted over to the backup data center
[15:53] and that is failover routing policy now
[15:56] the next policy we're going to talk
[15:57] about is geolocation and geolocation is
[16:00] really cool
[16:02] geolocation will route traffic based
[16:04] upon the location of your users so
[16:07] is your location routing
[16:09] i leave my house in florida i go visit
[16:13] my family and our village in greece
[16:15] i'm in my village in greece i go to
[16:17] connect the internet and something
[16:19] amazing happens
[16:20] i connect to the internet and i want to
[16:22] go to a page and it sends me to a page
[16:25] with greek writing all over it
[16:27] what happened is geolocation routing
[16:31] policy geolocation routing policy is i
[16:34] go to a country it looks at my source i
[16:37] p address by my source i p address it
[16:39] knows where i am where my location is
[16:42] and then it sends me to the closest
[16:44] website for example based upon my ip
[16:47] address so if in paris i'll get sent to
[16:50] a french website if i'm in the middle
[16:53] east i might get sent to an arabic
[16:54] website it is really cool figure out
[16:56] your source ip address figure out your
[16:58] source country and then route you to a
[17:01] different load balancer so for global
[17:03] organizations that might have an arabic
[17:04] webpage and a french web page and a
[17:06] spanish webpage an italian web page and
[17:08] a greek web page and a mandarin web page
[17:11] how neat is it that you can be able to
[17:13] figure out who the user is by their
[17:14] source ip address and then send them to
[17:16] the best page for them that is
[17:18] geo-location routing and it is cool
[17:21] now there is another geography-based
[17:23] routing option that you have and that's
[17:25] going to be called geoproximity routing
[17:28] and realistically speaking this you're
[17:30] going to use this when you want to route
[17:31] traffic based upon the location of your
[17:34] resources but have a little bit of
[17:36] control where you're sending your data
[17:38] so that's realistically speaking the
[17:40] geolocation routing policy it gives you
[17:41] the ability to modify the size of a
[17:43] region and shift your location your data
[17:45] from point a to point b
[17:47] now let's talk about some other much
[17:49] more commonly used ones latency based
[17:51] routing
[17:53] here's what latency routing is
[17:56] when the user goes to figure out what
[17:58] the website and it hits the dns server
[18:01] it will determine what region they're in
[18:03] and what's closest to them what has the
[18:05] lowest latency and it will send them to
[18:07] the environment that has the lowest
[18:08] latency which is going to give the best
[18:10] user experience
[18:12] the next one is called multi-value
[18:14] answer routing policy and the
[18:16] multi-value answer routing policy is
[18:19] kind of this random policy
[18:21] basically speaking you've got a couple
[18:23] of web servers and multi-value answer
[18:24] will just basically route randomly to
[18:26] whichever one it feels like at the time
[18:29] interesting
[18:30] random generally speaking if you can
[18:33] engineer things ahead of time it's
[18:34] generally a good thing so most
[18:36] organizations do not use the multi-value
[18:38] answer routing instead they have a
[18:39] policy but it's good to know that you
[18:41] have that option the last option we'll
[18:43] talk about and this is a great option is
[18:44] called weighted routing policy and
[18:47] weighted rallying and rated routing
[18:49] enables you to load share maybe you want
[18:51] to send 50 of your traffic to one place
[18:53] and 50 to another or 70 or 30 percent or
[18:57] better yet
[18:58] you've got your new your old website and
[19:00] it's running great everything's good
[19:02] users are happy
[19:04] your team makes a new website but you
[19:06] don't want to lose your old website so
[19:08] with dns you send 90 to your old one and
[19:12] 10 to the new website and you get
[19:14] feedback the new website works it's
[19:16] great it's great then you shift another
[19:18] 20 over the new website and then you
[19:19] shift everybody to the new website when
[19:20] you know it's good weighted routing
[19:23] perfect opportunity for new websites way
[19:25] to test new applications so
[19:28] aws dns flavor is route 53 we talked
[19:31] about simple routing we talked about
[19:32] fail over routing we talked about
[19:34] geolocation routing geoproximity routing
[19:38] latency-based routing multi-value answer
[19:40] routing and of course weighted router
[19:42] today we talked about the disaster
[19:43] recovery options we talked about data
[19:46] transfer speeds we talked about overhead
[19:48] on the link we talked about storage and
[19:50] why it's used on the cloud and we talked
[19:52] about dns routing policy
[19:54] thank you so much for watching this
[19:56] video i look forward to seeing you a new
[19:58] video very soon take care
[20:01] it was so nice having you join us for
[20:03] this video today let me tell you about
[20:05] some free services we do for the cloud
[20:06] community once per week we actually have
[20:09] a free
[20:11] question and answer session on live on
[20:13] youtube where you can come and ask us
[20:15] any questions you want about building
[20:16] your career related to cloud computing
[20:18] or networking we'll answer them in real
[20:20] time for you because we want to get you
[20:22] to your goals
[20:23] several more times per week we have
[20:26] guests from industry industry experts
[20:28] that i've known for decades that are
[20:30] movers and shakers that have changed the
[20:31] world that can give you information so
[20:33] you can build the best career i invite
[20:35] them periodically they are on my show
[20:38] if there's a chance to do some free
[20:39] training on our channel we'll do it live
[20:41] because we want you all to have the best
[20:42] skills for the best career so please
[20:45] subscribe and hit the bell i look
[20:47] forward to seeing you and i look forward
[20:48] to assisting you in your technology
[20:50] career thank you so much this is michael
[20:52] gibbs from go cloud architects