Intro to Cloud Computing

00:47:20
https://www.youtube.com/watch?v=UOtEkCNs940

Résumé

TLDRO vídeo aborda a computação em nuvem desde seus primórdios até as tendências atuais. Inicialmente, explica a evolução da computação, passando de centralizada para descentralizada, e retornando ao modelo centralizado com data centers e, por fim, à computação em nuvem. Descreve os três principais modelos de serviço: IaaS, PaaS e SaaS, e os tipos de nuvem: pública, privada e híbrida. Além disso, enfatiza a importância da segurança com um modelo de responsabilidade compartilhada, onde tanto o provedor quanto o cliente têm partes a gerenciar para manter o ambiente seguro. Há uma discussão sobre os principais provedores como AWS, Azure e Google, seus modelos de mercado e os desafios enfrentados, tais como segurança, conformidade, dependência de fornecedores e migração de dados. Os benefícios incluem flexibilidade, escalabilidade e economia de custos, enquanto problemas potenciais se concentram na segurança e na gestão de interfaces complexas. Também são destacadas tendências como Edge computing e a aplicação de inteligência artificial para melhorar serviços em nuvem.

A retenir

  • 🌥️ Computação em nuvem transformou a forma como as empresas utilizam a tecnologia.
  • 🔒 Segurança na nuvem requer atenção ao modelo de responsabilidade compartilhada.
  • 🌐 Principais provedores são AWS, Microsoft Azure e Google Cloud.
  • ⚙️ Modelos de serviço incluem IaaS, PaaS e SaaS.
  • 🚀 A nuvem oferece flexibilidade e escalabilidade.
  • 🌍 Tipos de nuvem: pública, privada e híbrida.
  • 🔄 Desafios incluem migração de dados e dependência de fornecedores.
  • 💡 Tendências emergentes: Edge computing e AI.
  • 💰 Modelos de preço na nuvem são geralmente mais acessíveis.
  • 📊 Inteligência artificial e machine learning são aplicados para melhorar serviços.

Chronologie

  • 00:00:00 - 00:05:00

    A introdução ao curso discute a definição e a história da computação em nuvem, seus benefícios, desafios como a segurança, e apresenta os principais fornecedores do setor, destacando a liderança da AWS, Microsoft e Google no mercado.

  • 00:05:00 - 00:10:00

    Explora o modelo de infraestrutura como serviço (IaaS), que oferece flexibilidade para gerenciar aplicativos em servidores na nuvem, abordando desafios de escalabilidade e manutenção durante picos de demanda, especialmente observados durante a pandemia.

  • 00:10:00 - 00:15:00

    O papel do cliente em manter a segurança em IaaS e os problemas comuns enfrentados durante a pandemia devido à migração apressada para a nuvem. Introduz serviço de plataforma como serviço (PaaS) e exemplo de ferramentas usadas para CI/CD.

  • 00:15:00 - 00:20:00

    O conceito de software como serviço (SaaS) é explorado, destacando como ele abstrai o hardware e fornece software via web. Discute tipos de nuvens: públicas, privadas e híbridas, explicando suas diferenças e casos de uso.

  • 00:20:00 - 00:25:00

    O modelo de responsabilidade compartilhada na segurança da nuvem é detalhado, mostrando como responsabilidades são divididas entre fornecedores de nuvem e clientes, com exemplos da AWS e Azure.

  • 00:25:00 - 00:30:00

    Examina a Cloud Security Alliance e seus controles de segurança, destacando a necessidade de avaliar a conformidade de novas empresas de nuvem com o framework CCM v4. Introdução aos conceitos gratuitos de camadas de nuvem para aprendizado individual.

  • 00:30:00 - 00:35:00

    Descreve os desafios de segurança, conformidade e lock-in do fornecedor na computação em nuvem, e como esses fatores influenciam decisões de migração e estratégia de TI em empresas nas Filipinas.

  • 00:35:00 - 00:40:00

    Benefícios como a capacidade de resposta em tempo real, análise e gestão facilitadas pela nuvem são discutidos, exemplificando com casos como veículos autônomos e precificação flexível, além de tendências como multicloud e edge computing.

  • 00:40:00 - 00:47:20

    Explora tendências futuras como servidores serverless e AI/ML na nuvem, ressaltando a flexibilidade que a nuvem oferece para implementações rápidas e escaláveis em várias indústrias.

Afficher plus

Carte mentale

Mind Map

Questions fréquemment posées

  • O que é computação em nuvem?

    Computação em nuvem é o uso de serviços e aplicativos hospedados remotamente, acessíveis via internet, ao invés de localmente em um computador pessoal.

  • Quais são os modelos de computação em nuvem mais comuns?

    Infraestrutura como Serviço (IaaS), Plataforma como Serviço (PaaS) e Software como Serviço (SaaS).

  • Quais são os principais provedores de computação em nuvem?

    Os principais provedores são a AWS, Microsoft Azure, Google Cloud, seguidos por outros como Alibaba, IBM, e Oracle.

  • Quais são os tipos de nuvem?

    Os tipos incluem nuvem privada, pública e híbrida.

  • Quais são os principais desafios da computação em nuvem?

    Segurança, conformidade, dependência do fornecedor, migração de dados e largura de banda.

  • O que é o modelo de responsabilidade compartilhada na nuvem?

    É um modelo onde o provedor de nuvem e o cliente compartilham responsabilidades pela segurança e manutenção do ambiente de nuvem.

  • Quais são os benefícios da computação em nuvem?

    Flexibilidade, escalabilidade, economia de custos, e um modelo de preço acessível.

  • O que é Edge computing?

    Edge computing é um paradigma de computação distribuída que traz o processamento de dados para mais perto dos dispositivos que geram os dados, reduzindo a latência.

  • Por que a segurança na nuvem é desafiadora?

    Devido à complexidade do ambiente em nuvem, com muitos serviços e interfaces que podem ter falhas de configuração se não forem geridos corretamente.

  • Quais são as tendências emergentes em computação em nuvem?

    Computação em múltiplas nuvens, Edge computing, inteligência artificial e machine learning.

Voir plus de résumés vidéo

Accédez instantanément à des résumés vidéo gratuits sur YouTube grâce à l'IA !
Sous-titres
fil
Défilement automatique:
  • 00:00:00
    [Musika]
  • 00:00:17
    Okay so Our first um topic for today is
  • 00:00:20
    intro to cloud computing um as
  • 00:00:23
    Uh you probably all have already used no
  • 00:00:26
    cloud computing One Way Or The Other or
  • 00:00:28
    even have deployed or maintaining it no
  • 00:00:29
    so this is more of a foundational
  • 00:00:32
    course so for for this morning no We'll
  • 00:00:35
    have Uh discussions on cloud computing
  • 00:00:37
    cloud computing security current
  • 00:00:40
    challenges Uh benefits and Trends to Uh
  • 00:00:44
    cloud computing and then we wrap it up
  • 00:00:46
    with the key learning points and some of
  • 00:00:48
    my references in in this course So what
  • 00:00:51
    is cloud computing So you h here the
  • 00:00:53
    cloud and this has been a common
  • 00:00:55
    password for quite some time move
  • 00:00:58
    everything to the cloud ba let's do
  • 00:01:00
    Cloud migration cloudy Secure so Uh
  • 00:01:03
    let's discuss What really is this Uh
  • 00:01:06
    Cloud that we're talking about okay to
  • 00:01:08
    dis spell then and demystify some of
  • 00:01:10
    this aspect so Uh Let's go back a bit to
  • 00:01:15
    Uh our computing history no so in the
  • 00:01:19
    early days no um We have a centralized
  • 00:01:22
    application only running on one isolated
  • 00:01:25
    computer I would say this is in the 60s
  • 00:01:27
    No when computers were starting to Uh Uh
  • 00:01:30
    pop up but of course the computers back
  • 00:01:31
    then Were A Whole floor no or even a
  • 00:01:35
    whole room No hence there're still just
  • 00:01:37
    starting to build up no and this was Uh
  • 00:01:40
    also the time where na transistors were
  • 00:01:42
    developed and microcontrollers no and
  • 00:01:45
    then came personal computing no this was
  • 00:01:47
    a most likely in the 80s And now we have
  • 00:01:51
    decentralized application so since you
  • 00:01:54
    have your own personal computing Uh
  • 00:01:55
    workstation you don't need a big one to
  • 00:01:58
    be able to process Uh applications and
  • 00:02:00
    and jobs no so each and every
  • 00:02:02
    workstation could Run independently okay
  • 00:02:05
    but Uh the main issue then was how do
  • 00:02:08
    you share data no How do you consolidate
  • 00:02:09
    no um well there's already computer
  • 00:02:12
    Networks in the 80s noo It's still in
  • 00:02:14
    the infancy stage no for common use no
  • 00:02:18
    and then came cloud computing so cloud
  • 00:02:20
    computing Uh when data Center start
  • 00:02:23
    booming no Uh likely in the 90s we have
  • 00:02:26
    this sort of Uh concept of cloud
  • 00:02:28
    computing but we weren't call it cloud
  • 00:02:30
    computing back then we're calling it
  • 00:02:31
    data centers and with cloud computing
  • 00:02:35
    you have centralized application Running
  • 00:02:36
    In data centers and then you're
  • 00:02:38
    connecting to it remotely either you
  • 00:02:40
    Submit your jobs and then you wait for
  • 00:02:42
    the results or you have an interface no
  • 00:02:44
    to that web application and then it's
  • 00:02:47
    being processed on the back background
  • 00:02:49
    and challenges of course for Central
  • 00:02:51
    application is Uh when running it close
  • 00:02:54
    to the users no either on the device or
  • 00:02:56
    the network itself No because um this
  • 00:02:59
    Edge device are the ones Generating data
  • 00:03:01
    and Uh if we had to submit it to a
  • 00:03:05
    Central location There are some pros and
  • 00:03:08
    cons back then so we'll see it in a bit
  • 00:03:11
    sa challenges Okay so These are the
  • 00:03:14
    major cloud computing infrastructure
  • 00:03:17
    players no sa from gartner and what we
  • 00:03:20
    could see here is Well of course aws
  • 00:03:23
    would be leading the Bunch no followed
  • 00:03:25
    closely by Microsoft and Google then we
  • 00:03:27
    have our Uh n layers and challengers no
  • 00:03:30
    from Alibaba Oracle IBM 10 Cent and
  • 00:03:34
    Huawei cloud no so for aws no this was
  • 00:03:40
    their Uh stat
  • 00:03:42
    now second quarter 2022 they own 30% 34%
  • 00:03:47
    of the market share of cloud computing
  • 00:03:49
    sa 200 billion Cloud market back then
  • 00:03:52
    then we have aure following closely with
  • 00:03:55
    21% and this is Uh as you could ano Yes
  • 00:04:01
    it's because it's Microsoft no and most
  • 00:04:04
    enterprise if not na well almost close
  • 00:04:07
    to all of them are running Microsoft Uh
  • 00:04:10
    Windows um hence it's just logical to
  • 00:04:13
    them no Uh to migrate let's say their
  • 00:04:16
    active Directory or iim to the cloud and
  • 00:04:20
    manage and maintained by Microsoft
  • 00:04:23
    because well Who's better to maintain a
  • 00:04:26
    an active Directory than the the creator
  • 00:04:28
    itself No but of course a ws has
  • 00:04:30
    something similar no ah Google Cloud
  • 00:04:32
    also has something similar no and then
  • 00:04:33
    we have again our our challengers no
  • 00:04:36
    from Alibaba IBM sales force 10 cents
  • 00:04:39
    and and Oracle now going to major cloud
  • 00:04:41
    computing model no we have here the
  • 00:04:43
    basic three no Uh there are still some
  • 00:04:45
    modif some nich types noo but Uh These
  • 00:04:48
    are the the common one we got
  • 00:04:50
    infrastructure as a service Where you're
  • 00:04:53
    basically given a a bear a bear computer
  • 00:04:57
    or an instance or a compute node no
  • 00:05:00
    where you are the one Uh who would be
  • 00:05:03
    managing the applications on it no Uh
  • 00:05:06
    and this somehow gives you flexibility
  • 00:05:08
    no ah since It's basically a Linux or a
  • 00:05:11
    Windows Server that's Uh in a data
  • 00:05:14
    Center somewhere Uh that you could
  • 00:05:16
    configure to your liking no and then you
  • 00:05:18
    could choose no your Resource do you
  • 00:05:20
    need more memory do you need more
  • 00:05:22
    storage do you need faster storage Uh do
  • 00:05:25
    you need How much CPU do you need So you
  • 00:05:27
    could configure this No it's just
  • 00:05:28
    basically Uh
  • 00:05:30
    uh what do you Actually need in your um
  • 00:05:34
    scenario No that that you're G to be Uh
  • 00:05:36
    the systems you're G to be deploying
  • 00:05:38
    um this adds flexibility to Uh
  • 00:05:43
    infrastructure managers No because used
  • 00:05:45
    to be before Yeah be having a cloud
  • 00:05:49
    vendor to subscribe to Uh you would
  • 00:05:51
    design your system you would do capacity
  • 00:05:53
    um prediction and typically you would
  • 00:05:56
    want to predict maybe 5 to 10 years
  • 00:05:58
    capacity no for that box So the problem
  • 00:06:01
    There Was well You need to pay and
  • 00:06:04
    purchase no Uh those hardware put it on
  • 00:06:07
    the data center and then on the initial
  • 00:06:09
    stages of your deployment You might not
  • 00:06:11
    have that much user or workload So you
  • 00:06:14
    probably maybe operating maybe 5% 10% at
  • 00:06:19
    most no for the first 1 to 2 years until
  • 00:06:22
    subscriber picks up No but Uh well
  • 00:06:24
    nowadays We have a thing called a system
  • 00:06:27
    or a service that goes viral no where
  • 00:06:30
    there's going to be an influx no of
  • 00:06:32
    users Uh in an instant and that would be
  • 00:06:35
    a problem No because let's say in in a
  • 00:06:37
    Year's time or in a few months time that
  • 00:06:38
    you deployed Uh you're maybe just using
  • 00:06:41
    Uh 51% no and then sa your your platform
  • 00:06:44
    got picked up no ah and Uh they want a
  • 00:06:48
    lot of users no want to subscribe to Uh
  • 00:06:50
    now you You might have a hard time um
  • 00:06:54
    Uh scaling that that platform in an
  • 00:06:57
    instant no and because now you have to
  • 00:06:59
    do Uh purchasing again I wait for the
  • 00:07:02
    delivery of the servers and then install
  • 00:07:04
    and configure it so that's that's going
  • 00:07:06
    to take some time and the demand might
  • 00:07:08
    not be there Uh by the time that Uh this
  • 00:07:11
    hardware gets delivered and that also Uh
  • 00:07:13
    highlights on the problem during the
  • 00:07:15
    pandemic I'm not sure if Uh Anyone Of
  • 00:07:17
    You are running a data centers or even
  • 00:07:20
    just a corporate um it infrastructure Uh
  • 00:07:24
    during the pandemic I had Uh deliveries
  • 00:07:27
    no Uh lead time as as long as almost a
  • 00:07:30
    year no the common um Uh lead time you
  • 00:07:34
    You heal from the vendor is 180 days no
  • 00:07:37
    and more Uh well because Uh There's a
  • 00:07:40
    shutdown Uh even the manufacturing is
  • 00:07:43
    has Slow down and logistics too So um if
  • 00:07:46
    you have a service during the pandemic
  • 00:07:48
    that you need uh some hardware
  • 00:07:50
    replacement or just typically you want
  • 00:07:52
    to Uh scale no Uh That's not going to be
  • 00:07:55
    possible and cloud computing boom Uh
  • 00:07:58
    especially during the pandemic Because
  • 00:07:59
    now you don't have to wait for anyone
  • 00:08:01
    you have your credit card sign up to aws
  • 00:08:04
    azure Google Cloud or Uh other no cloud
  • 00:08:07
    vendors and then you have your Uh Uh
  • 00:08:10
    instance no immediately for you Okay um
  • 00:08:14
    and again you could configure what you
  • 00:08:15
    want no And there's cost involved of
  • 00:08:17
    course
  • 00:08:18
    um but Uh Of course you already have
  • 00:08:21
    studied Iyung roi no for that so that's
  • 00:08:25
    infrastructure sa service Uh but do note
  • 00:08:27
    na You are the one maintaining the
  • 00:08:29
    softwares Uh and the security of the the
  • 00:08:33
    infrastructure sa service no so Uh we
  • 00:08:35
    got Uh cases where Uh also this this was
  • 00:08:39
    Uh a lot Uh this happened a lot during
  • 00:08:42
    the pandemic where again they cannot
  • 00:08:44
    purchase hardware they need to move to
  • 00:08:46
    Cloud because it's in It's instantly
  • 00:08:48
    there Uh and Thinking that cloud is
  • 00:08:51
    Secure No not realizing again that well
  • 00:08:53
    it is Secure as long as you You Secure
  • 00:08:55
    it Uh especially for infrastructure sa
  • 00:08:58
    service ah um hence Uh some of them Uh
  • 00:09:01
    got Uh security issues no and some of
  • 00:09:03
    them were breached no because of Uh the
  • 00:09:07
    wrong way of thinking how to Secure the
  • 00:09:08
    cloud next we have platform as a service
  • 00:09:11
    so platform as a service this one is Uh
  • 00:09:15
    well you have a provider that Uh Well
  • 00:09:17
    it's in the cloud no and you you would
  • 00:09:20
    Uh an example is a some some services
  • 00:09:23
    like managing and developing and testing
  • 00:09:26
    Uh applications along with it you have
  • 00:09:29
    your tools um through the platform you
  • 00:09:31
    run the test and development Uh I would
  • 00:09:34
    say sample of platform a service here
  • 00:09:37
    would be github docker and kubernetes no
  • 00:09:40
    so for mainly for continuous integration
  • 00:09:42
    and continuous deployment Uh Uh
  • 00:09:45
    workloads no so we got that and then
  • 00:09:48
    lastly we have saas no software sa
  • 00:09:50
    service here is well you as a subscriber
  • 00:09:53
    to SAS um you would not be the one
  • 00:09:57
    managing the back end no so It's again
  • 00:10:00
    It's software you're being abstracted to
  • 00:10:02
    the hardware on the back end um the
  • 00:10:05
    vendor would be Uh doing the patching
  • 00:10:07
    maintaining adding features No you just
  • 00:10:10
    use the system no either via a web
  • 00:10:13
    browser an app or an api no um So um
  • 00:10:18
    sample here would be maybe a Microsoft
  • 00:10:20
    365 Uh G Suite and Uh the rest no that's
  • 00:10:25
    a web based na system May um you could
  • 00:10:28
    also Uh plugin here Uh Spotify tidal for
  • 00:10:33
    Uh audio streaming service um because
  • 00:10:36
    well technically you're subscribing to
  • 00:10:37
    Uh the Stream No but and you are not
  • 00:10:40
    managing the platform you are not
  • 00:10:41
    managing the infrastructure so we could
  • 00:10:44
    consider them a software sa service okay
  • 00:10:47
    now for the types of cloud computing We
  • 00:10:50
    have a private Cloud so these are
  • 00:10:52
    applications that are hosted no on your
  • 00:10:54
    Server on the intranet within your
  • 00:10:57
    organization and with this No you are
  • 00:10:59
    responsible for managing Uh upgrading
  • 00:11:03
    updating the software system security
  • 00:11:05
    and other Resource no including Uh power
  • 00:11:08
    and cooling noo and and Space of course
  • 00:11:10
    um so there are still Uh organizations
  • 00:11:14
    that has Uh private Cloud Uh Actually
  • 00:11:17
    there's a lot of Philippine organization
  • 00:11:18
    that I would say have some private cloud
  • 00:11:21
    no and um they are using Uh technologies
  • 00:11:24
    because of course There's not Uh Uh you
  • 00:11:26
    can't deploy aws here ah It's on your
  • 00:11:29
    own no Uh Typical technology that you
  • 00:11:31
    would need no and deploying private
  • 00:11:33
    Cloud would be Zen ah
  • 00:11:36
    citrix also got VMware no um um and prox
  • 00:11:41
    mox That's a some technologies That You
  • 00:11:44
    might Uh want to Take a Look at if you
  • 00:11:46
    want to deploy your own Cloud
  • 00:11:47
    environment or
  • 00:11:49
    um yeah but again You're the one
  • 00:11:52
    maintaining it even the hardware youl Be
  • 00:11:55
    The One purchasing and Uh deploying it
  • 00:11:57
    um even the bandwidth no Uh It's on you
  • 00:12:01
    You have to um Uh subscribe to Uh telco
  • 00:12:05
    Uh ban no um and configure your network
  • 00:12:09
    and then you have your own hardware
  • 00:12:10
    firewall to protect your infrastructure
  • 00:12:13
    Okay so that's a private Cloud public
  • 00:12:16
    Cloud well This Is Where the provider no
  • 00:12:19
    Uh installs the software on the server
  • 00:12:22
    Okay basically it's it's the operating
  • 00:12:24
    system no ah and gives you access to
  • 00:12:26
    their interface ah and use the service
  • 00:12:28
    and and typically the provider would be
  • 00:12:30
    responsible for most of the management
  • 00:12:32
    on the back end So if there's any power
  • 00:12:35
    issue there's any bandwidth issue Uh
  • 00:12:38
    there's any cooling issue No that's
  • 00:12:39
    going to be part of Uh the provider no
  • 00:12:42
    and let's say There's a hardware failure
  • 00:12:45
    You're you're abstracted from that they
  • 00:12:47
    have redundancy They would be able to
  • 00:12:49
    recover with minimal or up to no
  • 00:12:52
    disruption
  • 00:12:53
    Actually and They're also doing the
  • 00:12:56
    upgrades no Uh and maintenance and
  • 00:12:58
    security to a degree no Uh again Uh
  • 00:13:01
    there is a what we called shared
  • 00:13:03
    responsibility model which we see in a
  • 00:13:06
    bit too now um since some scenarios no
  • 00:13:10
    Uh could not be solely deployed on
  • 00:13:13
    private Cloud or on the vice the reverse
  • 00:13:16
    cannot be solly deployed on the public
  • 00:13:17
    Cloud Uh we are seeing a lot of hybrid
  • 00:13:20
    Cloud So this is a mixture of both
  • 00:13:23
    private and public solutions where
  • 00:13:25
    You'll be in charge of handling and
  • 00:13:26
    managing this two Uh public and PR
  • 00:13:29
    private services together Uh and one of
  • 00:13:32
    the main um concerns here now is passing
  • 00:13:35
    the data in between this Cloud services
  • 00:13:37
    no and a lot of organizations here in
  • 00:13:39
    the Philippines are at this stage no
  • 00:13:41
    hybrid Cloud um especially the the the
  • 00:13:44
    older organizations noo most of startups
  • 00:13:47
    na you could see They're mayly in in
  • 00:13:50
    public Cloud or mga mga fintech and tech
  • 00:13:53
    startups so they would be they have
  • 00:13:56
    Cloud First Policy so they would deploy
  • 00:13:59
    everything in the Cloud and what they
  • 00:14:00
    have here Uh on prem would just be
  • 00:14:02
    workstation but for traditional
  • 00:14:04
    companies or those that have been there
  • 00:14:07
    Uh pre cloud computing days They they
  • 00:14:10
    they would have some sort of a hybrid
  • 00:14:12
    Cloud Where they have still their own
  • 00:14:14
    data centers and virtualization on prem
  • 00:14:16
    but they have solutions and workloads in
  • 00:14:19
    the public Uh Cloud Uh vendors hosted by
  • 00:14:23
    public clog vendors no and again um
  • 00:14:27
    issues that they would normally be
  • 00:14:29
    encountering is well sinking data
  • 00:14:31
    between two no Uh well there are
  • 00:14:34
    solutions and Uh controls in place no to
  • 00:14:36
    to figure out Which has the latest data
  • 00:14:39
    no and Uh do you lock Uh certain systems
  • 00:14:42
    no um to read only in case the sink
  • 00:14:45
    hasn't been Uh completed yet so Uh It's
  • 00:14:48
    It's pretty robust nowadays so There's
  • 00:14:50
    not much issue aside from well bandwidth
  • 00:14:53
    which also see in Okay so Uh Let's do Uh
  • 00:14:58
    si security so Cloud computer security
  • 00:15:01
    you probably heard this Uh tons of times
  • 00:15:04
    noo na cloud is secured so Let's move
  • 00:15:08
    our systems to the cloud Okay so there
  • 00:15:12
    is what we call a shared responsibility
  • 00:15:14
    model and this is the part where Uh a
  • 00:15:18
    lot of Uh the it guys are being
  • 00:15:21
    blindsided no Uh they thought that
  • 00:15:23
    everything here would be managed by aws
  • 00:15:26
    but
  • 00:15:27
    na since there's a lot of queries I I
  • 00:15:30
    guess to aws even in the early stages
  • 00:15:33
    where ba i moved to the cloud Why did we
  • 00:15:35
    get breach no Uh and aws released this
  • 00:15:38
    na Okay you got bridge because well
  • 00:15:40
    Officially based on our contract we're
  • 00:15:42
    Actually just maintaining the Uh orange
  • 00:15:45
    part above that would be your
  • 00:15:48
    responsibility as the customer so Uh
  • 00:15:51
    especially for easy2 or infrastructure
  • 00:15:54
    sa service right so network traffic
  • 00:15:55
    protection the encryption The integrity
  • 00:15:58
    of data Transit That's on you Uh the
  • 00:16:01
    file system no Uh there's any encryption
  • 00:16:03
    that you want to put on your Uh
  • 00:16:05
    operating system file system and Uh or
  • 00:16:08
    database no for that matter That's on
  • 00:16:10
    you um the firewall configuration ba If
  • 00:16:13
    You made it
  • 00:16:14
    Uh by mistake or whatever no to allow
  • 00:16:17
    any any then That's on you that's why
  • 00:16:19
    you got a breach no and the passwords no
  • 00:16:21
    identity access and Management to go Up
  • 00:16:24
    here even the patching of applications
  • 00:16:26
    no that's on you so if you have your
  • 00:16:28
    system with a password of password 123
  • 00:16:31
    then I don't think you could blame Uh
  • 00:16:33
    aws ready no If someone gets access to
  • 00:16:35
    your system because of that no um and
  • 00:16:38
    ultimately the customer data is your
  • 00:16:40
    responsibility so you do Uh risk
  • 00:16:43
    analysis no Uh if you deploy to anywhere
  • 00:16:46
    not only on cloud computing and
  • 00:16:47
    determine that if the system do get
  • 00:16:50
    breached no uh what kind of controls are
  • 00:16:53
    there to one Uh prevent them from
  • 00:16:55
    getting breach in the first place um
  • 00:16:58
    like maybe a firewall That's a
  • 00:17:00
    preventive control um antivirus would be
  • 00:17:03
    a detective control no in case something
  • 00:17:05
    goes wrong Uh AV should be able to
  • 00:17:08
    detect Uh any issues there no So you put
  • 00:17:12
    that in place and then after all the
  • 00:17:13
    control and the cost involved now you
  • 00:17:16
    would have a committee no to determine
  • 00:17:18
    if that is Uh a good investment no then
  • 00:17:21
    you decide if you want to move to the
  • 00:17:23
    cloud or any again any platform for that
  • 00:17:26
    matter okay now of course Uh since azure
  • 00:17:30
    is close second no Uh to aws they also
  • 00:17:33
    have this Uh shared responsibility model
  • 00:17:36
    but now they defined it a bit more
  • 00:17:39
    granular no than how aws designed it
  • 00:17:42
    Where they have some sort of I would say
  • 00:17:45
    marketing here no that says that okay we
  • 00:17:48
    have here the three different types of
  • 00:17:51
    cloud computing no sa aspa as ias and
  • 00:17:54
    you have your own prem and I would say
  • 00:17:56
    this is the some sort of the the the
  • 00:17:58
    marketing here that if you are on prem
  • 00:18:00
    you are responsible for everything so if
  • 00:18:03
    you want to transfer no the risk and
  • 00:18:05
    operations no to Uh a provider no to to
  • 00:18:09
    help you Uh minimize your cost on that
  • 00:18:11
    aspect Uh Yeah you could do I
  • 00:18:15
    infrastructure as a service where
  • 00:18:16
    physical host Networks and data Center
  • 00:18:18
    is Actually responsib the responsibility
  • 00:18:21
    of Microsoft No but ah as with Uh
  • 00:18:25
    aws you could see here that Uh os net
  • 00:18:28
    Networks applications no and data no Uh
  • 00:18:31
    that is within the control of the
  • 00:18:34
    customers so you can't really Uh pass
  • 00:18:38
    that no that responsibility to anyone Uh
  • 00:18:41
    It's on you and as you go to the left no
  • 00:18:44
    from platform a service software sa
  • 00:18:46
    service no Uh a lot of this is getting
  • 00:18:50
    maintained no by by the vendor So you
  • 00:18:52
    got less and less responsibility but
  • 00:18:54
    that responsibility is basically the
  • 00:18:56
    crown duels of your organization which
  • 00:18:58
    is
  • 00:18:59
    your accounts and identities and of
  • 00:19:01
    course your data and no cloud vendor
  • 00:19:04
    would be sane enough to get
  • 00:19:06
    responsibility for for those items
  • 00:19:10
    Especially if they don't have control on
  • 00:19:11
    how you configure your stuff
  • 00:19:13
    um like as mention if if your users use
  • 00:19:17
    weak password or They're not educated
  • 00:19:19
    enough to spot and detect fishing emails
  • 00:19:22
    Uh and they get compromised of course
  • 00:19:25
    actors would have access to your Uh data
  • 00:19:28
    at some point no so in regards to cloud
  • 00:19:32
    computing security No there is a a group
  • 00:19:34
    no called Cloud Cloud security alliance
  • 00:19:36
    no csa Uh that is Uh releasing no um a
  • 00:19:41
    framework for Uh cyber security control
  • 00:19:45
    in cloud computing and they are in
  • 00:19:46
    version 4 of the cloud control Matrix
  • 00:19:49
    ccm Uh v4
  • 00:19:52
    um it's not even Uh I would say a
  • 00:19:55
    hightech tool no but Uh it does cover no
  • 00:19:59
    Uh most aspects na you would see Uh that
  • 00:20:02
    cloud computing vendor should be
  • 00:20:03
    providing no and if you have a vendor
  • 00:20:06
    well most major vendors would be
  • 00:20:08
    compliant to this one no but and have
  • 00:20:10
    tweaks no for specific domains and
  • 00:20:13
    controls um I would say that Uh if you
  • 00:20:17
    are evaluating a cloud vendor especially
  • 00:20:19
    the new the new Commerce no and they
  • 00:20:22
    cannot Uh comply or have a control on
  • 00:20:25
    the 17 dom and 197 controls then you
  • 00:20:28
    might Uh want to reconsider no or again
  • 00:20:30
    Uh check out no is the cheaper um
  • 00:20:34
    implementation Uh in vendor a with less
  • 00:20:38
    security controls G to be Uh in the long
  • 00:20:41
    Run better than vendor B Which has all
  • 00:20:43
    this control but time 10 the price so
  • 00:20:46
    again there you would evaluate based on
  • 00:20:48
    your risk appetite which would be Uh
  • 00:20:51
    better so with this tool no ah you could
  • 00:20:54
    systematically assess no cloud
  • 00:20:55
    implementations ah and provide guidance
  • 00:20:58
    on which security control should be
  • 00:21:00
    implemented by which actor within the
  • 00:21:02
    cloud Uh supply chain and
  • 00:21:05
    to see no Uh the different Uh domains no
  • 00:21:10
    Uh we got audit and assurance
  • 00:21:13
    application and interface security a lot
  • 00:21:16
    more just not G to read this from and
  • 00:21:19
    then got here a good
  • 00:21:21
    um intro no to the cloud security
  • 00:21:24
    alliance Cloud control Matrix v4 from
  • 00:21:27
    NCC group with the link ah
  • 00:21:30
    below so ah in asia open Run course noo
  • 00:21:35
    we would be Uh requiring um the students
  • 00:21:39
    just to give you an overview then of
  • 00:21:41
    what are some of our activities no to
  • 00:21:44
    subscribe to Uh free cloud computing um
  • 00:21:47
    tier no so gcp aws and assure they have
  • 00:21:50
    free tiers No that well if this is the
  • 00:21:53
    first time you're hearing it and you
  • 00:21:54
    want to get your feet wet in cloud
  • 00:21:56
    computing you could subscribe to awg CP
  • 00:21:59
    sure and they will give you an instance
  • 00:22:00
    no Ah it's a micro instance Uh minimum
  • 00:22:03
    memory minimum CPU minimum storage but
  • 00:22:06
    at least you get to feel no How to log
  • 00:22:08
    in remotely how to use their Ui Uh so
  • 00:22:11
    again just getting your feet wet and
  • 00:22:13
    getting you Uh introduced na to cloud
  • 00:22:15
    computing and for our students no let's
  • 00:22:19
    say for they Set up an aws Fe tier
  • 00:22:23
    um This Would Be the the requirement no
  • 00:22:25
    So you go to aws free ah ws amazon.com
  • 00:22:29
    free then Uh unfortunately you're G to
  • 00:22:32
    be needing email address and credit or
  • 00:22:34
    debit card so
  • 00:22:35
    Uh there is a hit and Miss for
  • 00:22:38
    um maya or GCash Uh virtual credit card
  • 00:22:42
    but Uh in case you don't have your own
  • 00:22:43
    Uh Uh credit card no from Typical Uh
  • 00:22:47
    banks no Uh you could you could try out
  • 00:22:50
    no maya and Uh GCash for for this one
  • 00:22:53
    and
  • 00:22:54
    um you could have an micro instance on
  • 00:22:57
    aws for here with a public IP na you
  • 00:23:00
    could Uh login remotely do ssh install
  • 00:23:03
    some applications no um and Uh destroy
  • 00:23:07
    it no in case it goes wrong That Uh and
  • 00:23:09
    then recreate So this is a good again
  • 00:23:12
    This is some sort of Again one of their
  • 00:23:14
    marketing Uh tool no Uh of the cloud
  • 00:23:17
    vendors to help you get acquainted no to
  • 00:23:20
    their service to their interface to
  • 00:23:22
    their api buse well they gave you for
  • 00:23:24
    free Uh a year worth of a micro instance
  • 00:23:28
    but Uh you don't expect no to to run a
  • 00:23:31
    heavy usage website there Uh maybe some
  • 00:23:34
    a personal blog some Synchronization
  • 00:23:36
    server and just to try things out Uh
  • 00:23:39
    that will be good and of course if you
  • 00:23:40
    need more heavy duty and more more
  • 00:23:44
    capacity No you would most likely be
  • 00:23:47
    subscribing to Uh Um yung mga ano common
  • 00:23:51
    common Uh commercial Tears so next
  • 00:23:55
    section is challenges Uh into Uh Cloud
  • 00:23:58
    compu buing So you have here discussion
  • 00:24:01
    on five No security compliance vendor
  • 00:24:03
    lock in data migration and bwi so for
  • 00:24:08
    cloud no as mentioned that Uh when you
  • 00:24:11
    hear na cloud is Secure you Show them
  • 00:24:14
    your shared responsibility model So it
  • 00:24:17
    is Secure up to a certain degree and
  • 00:24:19
    then the rest would be your
  • 00:24:20
    responsibility So if you're not into Uh
  • 00:24:23
    securing the infrastructure on prem Uh
  • 00:24:26
    There's a chance that Uh
  • 00:24:28
    You might Uh not be the guy to be Uh the
  • 00:24:33
    one maintaining the security in the
  • 00:24:34
    Cloud because I I would tell you now
  • 00:24:36
    that Uh securing the cloud environment
  • 00:24:39
    It's much more complex than onprem in
  • 00:24:41
    onprem you have your perimeter once you
  • 00:24:44
    close off Uh typically you're Secure No
  • 00:24:47
    but that's Uh That's not how I would
  • 00:24:49
    design it no so even within the internal
  • 00:24:52
    space of Uh my onprem deployment we have
  • 00:24:55
    what we call zero trust so even those
  • 00:24:58
    machines within the infrastructure you
  • 00:25:00
    don't trust each other you Uh verify
  • 00:25:03
    each and everything that's G to be
  • 00:25:05
    requested Uh on each of components
  • 00:25:07
    Because You never know one of those box
  • 00:25:10
    could have been compromised and it's the
  • 00:25:11
    attacker doing the request on behalf of
  • 00:25:13
    that box So um
  • 00:25:17
    and I would say cloud computing is more
  • 00:25:20
    Complex to Secure Because there's a lot
  • 00:25:23
    of services there's a lot of interfaces
  • 00:25:26
    and a lot of tech is getting deployed as
  • 00:25:28
    we speak no
  • 00:25:29
    um the Ui You might have now might not
  • 00:25:32
    be the same Ui Uh by next week no um
  • 00:25:35
    because again there's a lot of
  • 00:25:36
    development and rnd in cloud that's also
  • 00:25:38
    being fed directly to the service
  • 00:25:40
    immediately Uh and if you are not Uh Uh
  • 00:25:46
    putting up to speed no on What's out
  • 00:25:47
    there no on What's the configuration
  • 00:25:49
    What's the features you'll get left out
  • 00:25:51
    And you'll leave your systems um Uh
  • 00:25:54
    vulnerable or an Exposed no So um like
  • 00:25:58
    on this Uh article no um Although Yeah o
  • 00:26:01
    almost a decade almost a decade back
  • 00:26:04
    where aws console bridge leads to Uh
  • 00:26:06
    demise of a service with appr proven
  • 00:26:08
    backup plan because well What happened
  • 00:26:10
    here was well they probably put all
  • 00:26:14
    their eggs in aws and even the backup
  • 00:26:17
    and someone got fished
  • 00:26:20
    um someone likely got fish in their Uh
  • 00:26:23
    organization and the Tet actor had
  • 00:26:25
    access to Uh code spaces coms Uh aws
  • 00:26:30
    console no and started Uh ruming it and
  • 00:26:33
    when they threat
  • 00:26:35
    actor um sense No that they were trying
  • 00:26:38
    to restore Uh they just deleted
  • 00:26:40
    everything so that's going to be a major
  • 00:26:44
    problem so dict they have a cloud First
  • 00:26:48
    Policy um this has been there since
  • 00:26:51
    Uh almost I think 4 five years Uh ago No
  • 00:26:54
    when i first heard of this where if
  • 00:26:56
    there's any government system no Uh They
  • 00:26:59
    would say na
  • 00:27:00
    okay Don't host it on prem Uh do a cloud
  • 00:27:03
    first no and ah them saying Cloud First
  • 00:27:06
    Policy does not necessarily mean Uh go
  • 00:27:08
    to aws immediately no so there's also at
  • 00:27:11
    the ict cloud no na you could subscribe
  • 00:27:13
    to okay but Uh now Of course you have to
  • 00:27:15
    do a cost benefit analysis no or um How
  • 00:27:19
    do you call this Uh Which one is more
  • 00:27:20
    appropriate on you and Uh there were
  • 00:27:24
    some um hesitant before with government
  • 00:27:27
    agencies moving to the cloud non because
  • 00:27:29
    of data sovereignity and and the likes
  • 00:27:31
    No but um we've also been seeing no a
  • 00:27:34
    lot of Uh government agencies having
  • 00:27:37
    their workload and email systems noo Uh
  • 00:27:40
    to the cloud say Uh there's a lot of
  • 00:27:44
    department Uh governmental department No
  • 00:27:46
    that has their emails in Uh Google
  • 00:27:49
    workspace and Office 365 so Uh they are
  • 00:27:53
    we are slowly adapting to cloud no Uh
  • 00:27:56
    And that's Uh mainly because
  • 00:28:00
    most likely when they did their cost
  • 00:28:02
    benefit analysis no
  • 00:28:04
    ah they realize that the the email
  • 00:28:08
    systems in the Cloud of Microsoft and
  • 00:28:10
    Google no Uh their technology on
  • 00:28:13
    filtering and uptime No they
  • 00:28:15
    are leaps and bounce better than our it
  • 00:28:19
    guy maintaining their own Uh email
  • 00:28:21
    server locally no so Uh again Uh in the
  • 00:28:25
    shared responsibility model a lot of
  • 00:28:26
    those patching and maintenances already
  • 00:28:28
    done no by the vendor and you're just
  • 00:28:31
    left no to maintain your Uh users no
  • 00:28:35
    identity and your customers data and
  • 00:28:38
    that made it a bit more simpler no to
  • 00:28:40
    decide on Why to move to the cloud
  • 00:28:43
    um and most most of our it personnels
  • 00:28:46
    here no Uh are already overworked as it
  • 00:28:50
    is no and now they'll have to do the
  • 00:28:52
    patching the testing and the maintenance
  • 00:28:54
    of this No that's Uh when you move to
  • 00:28:56
    cloud computing Uh email and environment
  • 00:28:58
    it gets uploaded Uh to them No they just
  • 00:29:01
    have to maintain the accounts now and
  • 00:29:03
    Sure no there's a brid some of most of
  • 00:29:05
    the system would be able to detect it
  • 00:29:07
    like Uh Maybe uh we from the Philippines
  • 00:29:10
    and then suddenly someone from Uh maybe
  • 00:29:14
    What um Nigeria logged it to your Uh
  • 00:29:18
    inbox Uh system would alert that it's
  • 00:29:20
    likely an impossible login okay and you
  • 00:29:23
    could get the it would get flug and
  • 00:29:25
    started doing start doing remedi
  • 00:29:28
    immediately on that account there's also
  • 00:29:30
    a Philippine specific Uh compliance no
  • 00:29:32
    for aws like sa for financial
  • 00:29:35
    Institution under bsp no there's the bsp
  • 00:29:38
    982 workload Uh which means well aws
  • 00:29:41
    will be compliant hence a lot of banking
  • 00:29:43
    systems are already being loaded to aws
  • 00:29:46
    and as Sure no because well they are
  • 00:29:48
    basically complying with bsp 982
  • 00:29:51
    circular 982 okay and other compliance
  • 00:29:54
    program no Uh if anyone would still ask
  • 00:29:56
    this is their Uh trophy cabinet no na um
  • 00:30:00
    aws is certified with
  • 00:30:03
    ISO gdpr compliant PCI Uh dss compliant
  • 00:30:08
    sock compliant and I would assume that
  • 00:30:11
    Uh our local onprem might not be Uh
  • 00:30:16
    compliant to Uh even one or two of this
  • 00:30:19
    noo So um but again they are compliant
  • 00:30:22
    on their infrastructure side no Uh
  • 00:30:24
    depending on your service you would need
  • 00:30:26
    to be the one Uh securing it
  • 00:30:29
    Uh in terms of security and privacy no
  • 00:30:32
    and you could also certify your
  • 00:30:34
    applications No that's not ah when you
  • 00:30:37
    put your applications in the Cloud let's
  • 00:30:39
    say in aws noo it does not necessarily
  • 00:30:41
    mean that you're automatically ISO
  • 00:30:43
    certified
  • 00:30:44
    Uh that the system is ISO certified the
  • 00:30:47
    infrastructure hosting it is ISO
  • 00:30:48
    certified but the software on it It's
  • 00:30:51
    going to be within your context So you
  • 00:30:52
    need to have your own Uh ISO Uh
  • 00:30:55
    certification compliance next problem we
  • 00:30:57
    have is well vendor luckin So if Uh You
  • 00:31:01
    are a main user of aws um It's G to be a
  • 00:31:05
    bit hard for you to move to Uh digital
  • 00:31:08
    ocean or Oracle or azure Nono and I
  • 00:31:11
    would say as a company no Uh each and
  • 00:31:15
    every one of them they would want to
  • 00:31:16
    retain their customer so one of their
  • 00:31:18
    retention mechanism is They would put in
  • 00:31:21
    a lot of Services na you could subscribe
  • 00:31:23
    to and make it Easy for you And they
  • 00:31:26
    have a lot of Services na you could
  • 00:31:28
    subscribe to that might not be available
  • 00:31:29
    on other um cloud providers no and Uh
  • 00:31:33
    They're tooling no um and say the the
  • 00:31:36
    command line and api of course only Uh G
  • 00:31:40
    to be working no on their um platform so
  • 00:31:44
    there are middleware solutions that Uh
  • 00:31:47
    translates no Uh to multiple Cloud
  • 00:31:49
    vendor where Uh you have an api that I
  • 00:31:52
    want to create an instance and then you
  • 00:31:54
    could choose No I want this on azure or
  • 00:31:56
    aws or node or Google Cloud and it does
  • 00:31:59
    the translation for you and connects to
  • 00:32:01
    those services and instantiate for that
  • 00:32:03
    so that's Uh mainly um for no multi
  • 00:32:08
    Cloud environment no and I would say
  • 00:32:10
    doing multicloud is Actually Uh
  • 00:32:14
    Uh not not only better no Uh It Might Be
  • 00:32:18
    necessary again You don't want to put
  • 00:32:20
    your eggs on one basket like what
  • 00:32:22
    happened with Uh code spaces Uh a while
  • 00:32:25
    ago
  • 00:32:26
    um but lessons learned I mean this Cloud
  • 00:32:30
    vendors also Uh assess no their security
  • 00:32:34
    and the lessons learned that they are
  • 00:32:35
    encountering on various issues of their
  • 00:32:37
    customers and they are putting up Uh
  • 00:32:40
    solutions for it no let's say Uh for
  • 00:32:44
    azure no even if you delete emails no Uh
  • 00:32:47
    it's it's still being held no for quite
  • 00:32:49
    some time no for an admin to be able to
  • 00:32:52
    recover in case Uh It's a track After
  • 00:32:55
    deleting this Uh fishing emails to hide
  • 00:32:57
    their their tracks no so Uh or aws being
  • 00:33:00
    Uh S3 buckets being deleted No there are
  • 00:33:03
    now solutions that you could make it
  • 00:33:05
    that it could retain it for quite some
  • 00:33:08
    time after decision Uh for restoration
  • 00:33:11
    Just in case so they are Uh Uh
  • 00:33:15
    researching they are doing their lessons
  • 00:33:17
    learned and they are feeding it back to
  • 00:33:19
    their
  • 00:33:20
    services no next issue we have is Uh
  • 00:33:25
    data migration So you want you want to
  • 00:33:28
    move to the cloud and you have let's say
  • 00:33:30
    one petabyte of information in your
  • 00:33:33
    onprem Uh service no imagine moving that
  • 00:33:37
    with the bandwidth and the cost of
  • 00:33:39
    bandwidth here in Uh the philippines no
  • 00:33:41
    Uh moving it to the cloud would take
  • 00:33:45
    some time um and I got this um issue
  • 00:33:51
    similar to that they want to move the
  • 00:33:52
    data to the cloud and when they computed
  • 00:33:55
    it's going to take them the months to
  • 00:33:57
    migrate all the Data with their current
  • 00:33:58
    bandwidth and if they upgrade their
  • 00:34:00
    bandwidth well telco Uh
  • 00:34:05
    um problem here no Uh is the cost so we
  • 00:34:08
    have Uh one of the highest cost of
  • 00:34:11
    bandwidth na in Uh Asia no um and that
  • 00:34:16
    entails no um you would need that band
  • 00:34:19
    wid to be able to move to the cloud and
  • 00:34:22
    secondary once all your data is in the
  • 00:34:25
    Cloud you would have to retrieve it one
  • 00:34:27
    way or the The Other no for processing
  • 00:34:28
    the on onprem on your workstations no So
  • 00:34:32
    you need bandwidth back and
  • 00:34:35
    um some Cloud providers charge bandwidth
  • 00:34:39
    based on the directionality ah of course
  • 00:34:41
    there's also cost to them so if you are
  • 00:34:44
    a heavy usage of their network you're G
  • 00:34:46
    to get charged at some point and a lot
  • 00:34:48
    of um deployment that I have encountered
  • 00:34:51
    did not consider This no the cost of the
  • 00:34:53
    Band with na here in the Philippines and
  • 00:34:56
    the cost of the Band with in the Cloud
  • 00:34:58
    only to realize in computing it No there
  • 00:35:01
    Uh supposedly return on investment did
  • 00:35:03
    not make sense after 6 months because
  • 00:35:05
    their system uses a lot of bandwidth and
  • 00:35:08
    let's say on the case of aws uploading
  • 00:35:11
    to aws is free because they want your
  • 00:35:13
    data downloading your data is another
  • 00:35:16
    issue They're G to charge you per
  • 00:35:17
    Gigabyte so again depends on how you
  • 00:35:20
    would Take a Look at that no So what are
  • 00:35:24
    the the benefits of
  • 00:35:26
    Uh computing no we got futuristic use
  • 00:35:29
    case no sa especially with Uh open Run
  • 00:35:32
    no and 5g Uh and autonomous Uh vehicles
  • 00:35:35
    No they need hybrid ecosystem Uh you
  • 00:35:38
    have host that process data close to the
  • 00:35:41
    end points for realtime responsiveness
  • 00:35:43
    because again if you have a self-driving
  • 00:35:45
    car you don't want to have that
  • 00:35:47
    self-driving car C no a person crossing
  • 00:35:50
    The Road while on let's say Uh 60 km per
  • 00:35:53
    hour Uh um Run no Uh and have that that
  • 00:35:58
    sensor no as the cloud no should i break
  • 00:36:00
    Uh or should I just run him over no um
  • 00:36:04
    you don't want that you have you want
  • 00:36:06
    need to have that Uh responsiveness in
  • 00:36:09
    real time and you cannot have it decided
  • 00:36:12
    on the cloud because well you have
  • 00:36:14
    latency in the cloud no Uh that 60 km
  • 00:36:17
    per h and someone crossing Uh
  • 00:36:18
    milliseconds count you don't want to Uh
  • 00:36:21
    have that delay and well worst case what
  • 00:36:24
    if suddenly your car gets disconnected
  • 00:36:27
    Uh for one reason or The Other and now
  • 00:36:30
    the car cannot decide whether to hit the
  • 00:36:32
    brakes or Uh just proceed so those type
  • 00:36:36
    of Uh
  • 00:36:39
    um systems no has their own Uh Central
  • 00:36:44
    um um decision making within the the the
  • 00:36:49
    equipment itself noo and have the data
  • 00:36:51
    Uh telemetry data for for that matter
  • 00:36:54
    and some other updates no or um
  • 00:36:58
    information sent to the central Cloud
  • 00:37:00
    for correlation and analytics um so Uh
  • 00:37:04
    another is Again the amount of data
  • 00:37:06
    being generated would make transmitting
  • 00:37:08
    this No and storing this centrally going
  • 00:37:12
    to be a major issue again we discussed
  • 00:37:15
    security latency responsiveness
  • 00:37:17
    analytics management and governance but
  • 00:37:20
    imagine a CCTV operations no Uh where
  • 00:37:25
    you have a cloud Uh storage um that
  • 00:37:29
    might work at home for 4 f si cameras
  • 00:37:32
    but even then you might need maybe 100
  • 00:37:34
    Mbps bandwidth na for a high def or a 4k
  • 00:37:37
    resolution camera now imagine having a
  • 00:37:40
    CCTV solution for a building that has
  • 00:37:43
    a 200 cameras that band wid is G to be
  • 00:37:47
    massive so Uh you typically would want
  • 00:37:50
    to have a hybrid like a hybrid Cloud
  • 00:37:53
    have those data stored Uh locally and
  • 00:37:56
    then you have some detections or alerts
  • 00:37:58
    no being sent to the cloud no as a Uh
  • 00:38:01
    for Central analytics and Just in case
  • 00:38:03
    Maybe uh KN on wood no the building
  • 00:38:05
    burnt down No you have still some Uh
  • 00:38:08
    information that was sent to the cloud
  • 00:38:09
    before Uh everything went to dust um you
  • 00:38:12
    also would Uh need noo Uh as a benefits
  • 00:38:14
    for cloud computing is Uh streamline
  • 00:38:17
    analytics Uh where um again with a lot
  • 00:38:22
    of this data from the edge Uh endpoints
  • 00:38:25
    noo Uh being sent to the cloud you got
  • 00:38:28
    numerous Uh analytical Uh advantage no
  • 00:38:31
    Uh when you could do some processing on
  • 00:38:34
    this data sets Uh that are structured
  • 00:38:37
    sem structured on structured no and you
  • 00:38:39
    get some insights no um This Is Where
  • 00:38:42
    big data comes in Uh you got all this
  • 00:38:44
    information now you could try to predict
  • 00:38:46
    no Uh market
  • 00:38:48
    Trends buying patterns of consumers um
  • 00:38:53
    or Uh information let's say related to
  • 00:38:56
    traffic so some this Uh Uh are Actually
  • 00:38:59
    being used right now Uh let's say if
  • 00:39:01
    you're a user of Uh Google maps or ways
  • 00:39:05
    for traffic um they are using Uh and
  • 00:39:08
    streaming streamlining analytics from
  • 00:39:10
    various sources that's the devices
  • 00:39:12
    feeding to the cloud that well this
  • 00:39:15
    Street is congested because Based on my
  • 00:39:18
    Uh learning based on GPS coordinates
  • 00:39:20
    there are Uh 50 cars with way here
  • 00:39:23
    traveling at 3 to 4 km per Cloud Uh
  • 00:39:28
    service knows that no and feeds it to
  • 00:39:30
    the other Uh users no let's say of ways
  • 00:39:34
    and Uh rerouting them No not Uh
  • 00:39:36
    depending on their destination that
  • 00:39:37
    don't use this Street or this Uh Highway
  • 00:39:40
    because it's congested and they would be
  • 00:39:41
    able to rout you so that's analytics
  • 00:39:44
    that work for us um benefits of cloud is
  • 00:39:48
    also flexible pricing models because
  • 00:39:50
    well Uh as with Uh your onprem no where
  • 00:39:54
    you have to purchase everything and then
  • 00:39:55
    you have to wait for it Then you have to
  • 00:39:57
    set it up no um with cloud computing
  • 00:40:00
    It's mostly opex based no So you don't
  • 00:40:03
    have to wait for it It's instant Uh
  • 00:40:05
    subscribe to whatever instance or
  • 00:40:07
    service you need with the corresponding
  • 00:40:09
    charge of course Uh you would be able to
  • 00:40:12
    use it and um with this No again You
  • 00:40:15
    don't have to
  • 00:40:17
    um instantiate na all the the the
  • 00:40:20
    Resource you need for the next 5 years
  • 00:40:23
    you could even make it Uh elastic no
  • 00:40:26
    where that's Say You are a seasonal
  • 00:40:29
    um system where you only have Uh a huge
  • 00:40:33
    amount of traffic during
  • 00:40:35
    Uh sale no let's say you are the likes
  • 00:40:38
    of Lazada or shopee um and then Uh after
  • 00:40:41
    the sale noo um where you have let's say
  • 00:40:44
    10 million customers you need to scale
  • 00:40:47
    up immediately um and then after the se
  • 00:40:49
    no the the number of customers Dies down
  • 00:40:52
    to maybe a million or 2 million no and
  • 00:40:54
    you have All These compute resources if
  • 00:40:56
    you are on prem you're not actually
  • 00:40:58
    utilizing no with cloud computing No you
  • 00:41:00
    could downsize it na Uh within within a
  • 00:41:03
    few minutes No that Uh No one's using
  • 00:41:05
    and if suddenly There's a spike right
  • 00:41:09
    the cloud computing Uh solution no would
  • 00:41:11
    be able to scale up no and Uh produce Uh
  • 00:41:15
    cloned instances of your system and
  • 00:41:17
    enabling it to handle Uh large usage no
  • 00:41:20
    and you could do this automatically or
  • 00:41:21
    on schedule so and that makes your cost
  • 00:41:24
    na more flexible too and of course so
  • 00:41:27
    with the compliance no there's hassle fa
  • 00:41:29
    to compliance Because if they ask na is
  • 00:41:31
    your data Center ISO certified No you
  • 00:41:34
    you saw the the badges of aws a while
  • 00:41:36
    ago Uh they are um at least compliant to
  • 00:41:40
    those Uh frameworks then lastly we have
  • 00:41:43
    Uh Trends no to to cap up the cloud
  • 00:41:47
    computing course so for Trends we got
  • 00:41:51
    multicloud Edge computing serverless ai
  • 00:41:53
    and machine learning so at least the top
  • 00:41:55
    two we have already some discussions So
  • 00:41:57
    we're
  • 00:41:57
    multicloud refers to again Uh you have
  • 00:42:00
    various vendors no Uh in cloud computing
  • 00:42:04
    that Uh You are using Uh for your Uh
  • 00:42:07
    different use cases no so I've seen
  • 00:42:09
    organizations subscribe to azure because
  • 00:42:13
    they're mainly operating in Microsoft um
  • 00:42:16
    environment no so especially for their
  • 00:42:19
    active Directory no in the Cloud Uh and
  • 00:42:22
    their email too No but they have their
  • 00:42:23
    workloads in aws no so that that's
  • 00:42:26
    possible um and with that they are also
  • 00:42:30
    doing a cross backup I've seen this work
  • 00:42:32
    na where Uh data backups in azure while
  • 00:42:36
    they have their own backups within azure
  • 00:42:38
    They would ship it to aws and They were
  • 00:42:40
    Uh their justification was in case
  • 00:42:43
    something goes wrong with azure no which
  • 00:42:45
    highly unlikely No but in case something
  • 00:42:47
    goes wrong or their infrastructure gets
  • 00:42:50
    breached in azure they would still have
  • 00:42:52
    their own isolated is copy on Uh aws and
  • 00:42:56
    vice versa
  • 00:42:57
    They're They're backing up their aws
  • 00:43:00
    infra within aws and also within azure
  • 00:43:03
    so in case anything happens to their aws
  • 00:43:05
    instance They would be able to recover
  • 00:43:06
    from azure backup o but ah imagine the
  • 00:43:10
    cost of Uh those Synchronization No it's
  • 00:43:12
    a lot No but they did their Math And
  • 00:43:15
    they said that it works for them so yeah
  • 00:43:18
    I guess it works for them um next is a
  • 00:43:21
    trend for Edge computing We have a
  • 00:43:23
    separate course noo Uh for Edge
  • 00:43:25
    computing um as fundamental I think we
  • 00:43:28
    have it next week So I'll also be
  • 00:43:30
    lecturing for Edge computing but Yeah
  • 00:43:32
    this is a distributed computing paradigm
  • 00:43:34
    so imagine right Uh where it used to be
  • 00:43:37
    centralized um in the early days of
  • 00:43:40
    computing Now they Uh decentralize it
  • 00:43:43
    with personal computing and centralized
  • 00:43:45
    it back for cloud computing Now we're
  • 00:43:49
    trying to decentralize it again using
  • 00:43:51
    Edge computing so it's Uh like a a
  • 00:43:54
    rubber band No but um there's some
  • 00:43:57
    similarities and also differences No
  • 00:44:00
    because with Edge computing you have
  • 00:44:01
    good computing power Uh on your Edge
  • 00:44:04
    devices but for the most part you have
  • 00:44:06
    Uh issues on power consumptions It's
  • 00:44:10
    most likely battery operated for iot and
  • 00:44:12
    then you have network traffic to
  • 00:44:14
    consider to transmit those data to the
  • 00:44:17
    centralized cloud So um Again the
  • 00:44:21
    premise of Edge computing is there are
  • 00:44:23
    some issues with Cloud that Uh you
  • 00:44:26
    cannot just because of Physics and I
  • 00:44:28
    would just say latency and bandwidth you
  • 00:44:31
    need bandwidth you need and there's
  • 00:44:32
    going to be latency
  • 00:44:34
    ah for now we could only do Uh maximum
  • 00:44:38
    speed of light but even speed of light
  • 00:44:39
    they have some latency
  • 00:44:41
    um so Uh We want as with the similar to
  • 00:44:45
    autonomous vehicle scenario We want
  • 00:44:47
    those decision to be happening within Uh
  • 00:44:50
    the environment of that iot especially
  • 00:44:52
    for realtime decision making that could
  • 00:44:55
    cause Harm No Uh for any individual So
  • 00:44:59
    that's Uh how Edge computing is Uh
  • 00:45:02
    coming into play okay now um so we got
  • 00:45:07
    three types no of flug computing no Uh
  • 00:45:10
    we now Have a fourth onea a lot more No
  • 00:45:12
    but this one is a gaining traction
  • 00:45:15
    function na sa service where you could
  • 00:45:17
    develop your code no Uh and run this
  • 00:45:19
    applications without having to manage
  • 00:45:21
    the infrastructure you don't need an eas
  • 00:45:23
    to instance you don't need a server you
  • 00:45:25
    would just create a function and any of
  • 00:45:29
    your apps or system needs that function
  • 00:45:32
    um aws lambda Google Cloud functions
  • 00:45:36
    Microsoft azure functions had that had
  • 00:45:38
    those systems ready to serve no Uh your
  • 00:45:41
    function and return the results on
  • 00:45:44
    Whatever your user needs based on how
  • 00:45:46
    you code your functions of course and if
  • 00:45:48
    you need to scale this case let's say
  • 00:45:50
    again There's a sale and a lot of users
  • 00:45:52
    are hitting this functions the scaling
  • 00:45:55
    would be done automatic for you but do
  • 00:45:57
    note No that while it The scale No it is
  • 00:46:01
    Uh per request You need to pay for the
  • 00:46:03
    actual usage so the more request the
  • 00:46:06
    more Uh Uh charges gets to your credit
  • 00:46:10
    cards there's also ai and machine
  • 00:46:12
    learning know lastly for cloud computing
  • 00:46:16
    because well um while you could have
  • 00:46:19
    your own ai and mL platforms Uh in your
  • 00:46:22
    office Uh a lot of this platforms
  • 00:46:25
    requires gpus Uh which is hard to come
  • 00:46:28
    by because of crypto mining Uh a lot of
  • 00:46:32
    the cloud vendors already have them Uh
  • 00:46:35
    available for you you just have to
  • 00:46:36
    subscribe to their instances with gpus
  • 00:46:39
    Uh and then Run your ai and machine
  • 00:46:41
    learning models on them Uh could do
  • 00:46:45
    natural language processing Uh computer
  • 00:46:48
    vision speech recognition so Uh those
  • 00:46:50
    are just some of the types of solutions
  • 00:46:53
    that you could Run on cloud computing Uh
  • 00:46:56
    is
  • 00:46:57
    with
  • 00:46:58
    gpus okay okay and that waps about our
  • 00:47:02
    cloud computing got here some key
  • 00:47:04
    learning points
  • 00:47:12
    [Musika]
Tags
  • Computação em Nuvem
  • Segurança
  • Modelos de Serviço
  • Provedores de Nuvem
  • Desafios
  • Tendências de Mercado
  • IaaS
  • PaaS
  • SaaS
  • Edge Computing