cloud essentials 3

Today let's talk about multicloud.

Multicloud is a cloud computing approach made up of 2 or more cloud environments.

Let's get started.

So, we've got a public environment here as well as a private cloud.

There's an important distinction to be made here: multicloud is not exactly the same thingas hybrid cloud.

distinction : 1. (특히 비슷하거나 관련이 있는 것들 사이의 뚜렷한) 차이[대조] 2. 뛰어남, 탁월함

Hybrid cloud implies that your workloads a working together across multiple clouds - so
interoperability and portability of your workloads.

interoperability : 사용자가 각각의 컴퓨터를 사용하여 네트워크로 접속된 모든 시스템 자원을 자유롭게 이용할 수 있는 것. 각각의 컴퓨터에서 다른 컴퓨터가 관리하는 데이터로 액세스를 할 수 있거나 다른 컴퓨터에 있는 프로그램을 기동하여 서비스를 받을 수 있다.
portability : 이식성, 원시 프로그램을 다른 기종으로 옮기는 것이 얼마나 용이한가를 나타내는 정도. 다른 기종에 다시 컴파일하기만 해도 바로 돌아가는 프로그램은 이식성이 높은 것이다.

Multicloud doesn't have that same requirement, but we are seeing that, you know, a lot of enterprise users and customers are using multi and hybrid cloud strategy together. 

I'd say that the growth of containers and Kubernetes technology has really enabled the
growth of multicloud.

Production-Grade Container Orchestration

So, we're seeing almost every big cloud provider out there providing manage Kubernetes, as well as it being used on prem in the private sector to help modernize legacy(유산) applications.

So why would you want to use multicloud and the strategy for your cloud computing?
Well, there's a number of reasons.

For one, let's say that you want to hit that big three-nines number as far as (~에까지) availability goes.

Multicloud strategy can help you hit that by ensuring that, you know, it's one thing when the clouds of one of your application workloads stops to work you still have another cloud that's supporting your applications.

In addition, you can enable better user experience.

So, let's say you have users on different sides of the globe.

By routing them to the cloud that's nearest to them, you can ensure lower latency and
better user experiences.

 

route : (특정한 루트를 따라 무엇을) 보내다[전송하다]

latency : 숨어 있음, 보이지 않음; 잠복, 잠재

Finally, what about specific integrations that only work on a particular clouds?

Say you have some sensitive firewall data that you don't want to put in the public cloud,

but you need to build integrations for.

By taking advantage of a multicloud approach, you can build workloads on the private side

that can take advantage of that private, sensitive, customer data.

I'd say there's 3 major ways and things that you want to tackle when taking on a multicloud strategies, and I'm gonna introduce these as three pillars.

 

pillar : 1. (다리·건물 지붕 등을 받치는, 특히 장식 겸용의 둥근) 기둥 2. (둥근 기둥 모양의) 기념비[물] (=column)

The first one is gonna be automation.

So, to do this I, let me introduce kind two different types of users: so, we've got Ops,

as well as Devs.

So, operations engineers are going to want to be able to spin up new clusters as well

as manage and, kind of, see the different clusters that have been created.

So, they're gonna want to take advantage of a unified, kind of, dashboard to do something like that.

So, we'll have the control plane here.

Let's say the ops engineers want to spin up two Kubernetes clusters.

So, they'll essentially figure out that configuration, and then go to the control panel, and tell it to spin up those clusters.

At that point you know they can specify where they want those clusters to be created.

Let's say they spin up two.

Kubernetes clusters, one in each of the clouds.

On, the other side with devs, let's say that they've built out a new application, a Docker

container, and they're ready to push it.

So, they've got a container built, and they wanted to deploy it the same way to both of
the clouds but you know it can be quite difficult if they didn't have a control plane to figure out the different authentications and configurations of the unique different clouds and kind of deploy that both times.

 

authentication : 입증, 증명; 인증

configuration : 1. 배열, 배치; 배열[배치] 형태 ,2. 환경 설정

In addition, this really gets in the way of traditional kind of CICD and DevOps workflows.

So, by taking advantage of that control plane, they can spin up that container in both of

the clouds with a single command.

Say that it's application 'A'.

So, I think that really hits this requirement for automation with multicloud.

The next kind of requirement that I want to hit on here is visibility.

Visibility is very important for working with multiple clouds because, you know, they each

have their own unique dashboards but to have a single unified way of managing those things is very important.

So, for operations engineers, by taking advantage of a single control dashboard, they can have access to a unified kind of approach and see all the different, you know, clusters that they've spun up and even drill down deeper and see things like pods and deployments in applications across multiple clouds.

 

unify : 통합하다, 통일하다

The last thing I want to touch on, the last pillar, is governance.

These days regulatory and compliance policies are becoming increasingly strict and they
differ from geography to geography.

 

regulatory : (산업·상업 분야의) 규제력을 지닌

Let's say the operations team has a compliance policy that they want to push out across all of these clouds.

So, first they'll have to, you know, write out that compliance policy, and then, you know, if they didn't have access to a control plane, this can be a very kind of painstaking(공들인),
or rather time-consuming, approach.

So, by taking advantage of the control plane they can push that compliance policy across
multiple clusters with a single command and ensure those security policies for compliance are being met.

I'd say that these are the 3 major things that you want to tackle when taking on a multicloud strategy.

Multicloud has been a growing focus for us here at IBM.

If you want to learn more about our multicloud management strategy, check out some of the links in the description below.