IT complexity simplification framework

Summary

このビデオスクリプトでは、Red HatのグローバルチーフアーキテクトリーダーであるナダンとIBMフェローでCTOのカイル・ブラウンが、IT環境の複雑さを簡素化する物語を語っています。彼らはIBMの多様なプラットフォームと環境を例に、自動化戦略、コンテナ化、データ管理、アプリケーションのモダン化、開発者ツール、そしてAIの取り入れ方について詳述しています。これらの技術的アプローチは、企業の効率化と標準化を促進し、ビジネス価値を実現する上で重要な役割を果たします。

Takeaways

• 🌐 タスクの多様性: IBMはZプラットフォーム、クラウド、オンプレミス、仮想化、エッジコンピューティングなど、多様な環境でワークロードを管理しています。
• 🤝 IBMとRed Hatの協力: 両社は共通のソリューションとアプローチを通じて、お客様のための生活をより良いものにするための取り組みを進めています。
• 🔄 自動化戦略の重要性: 多様な環境を管理するためには、人を増やすのではなく、自動化された管理が不可欠です。
• 📦 コンテナ化戦略: ワークロードの類似性を見つけ、共通のイメージとコンテナアプローチを通じて、異なるプラットフォーム間のワークロードを類似させています。
• 🔄 仮想化からコンテナ化へのシフト: オープンソースの進化と仮想化のコストの増加により、より多くのコンテナ化環境への移行が進んでいます。
• 📈 データの3つの形態: 静止データ、移動データ、非構造化データ/コンテンツを効果的に管理し、AIの燃料として活用する必要があります。
• 🛠️ アプリケーションのプラットフォーム: コンテナ化はアプリケーションをサポートし、カスタマイズされたアプリケーションや既存のアプリケーションを活用する基盤を提供しています。
• 👨‍💻 開発者ツールの強化: CI/CD環境の爆発的増加に対処し、分散環境でより効果的に働くためのツールと標準を提供しています。
• 🤖 AIの役割: AIはコードの書き方への助けとなるパートナーであり、開発者による正確な仕様とテストの定義が依然として必要とされています。
• 🛡️ セキュリティの重要性: セキュリティはすべてのレイヤーにわたって重要な要素であり、管理、倫理、セキュリティの問題を回避するためには常に考慮する必要があります。

Q & A

ナダンはRed Hatでどのような役職を務めていますか？

ナダンはRed Hatのグローバルチーフアーキテクトリーダーであり、CTO組織を率いている。

ケール・ブラウンはIBMのどのようなポジションにいますか？

ケール・ブラウンはIBMフェローで、IBM CIOオフィスのCTOを務めている。

ITの複雑さを簡素化するという物語とはどのようなものですか？

ITの複雑さを簡素化する物語とは、ITの技術的地形の複雑さを解消し、より効率的で管理しやすいIT環境を作り出すプロセスを指す。

IBMが抱えているプラットフォームと環境はどのようなものでしょうか？

IBMはZプラットフォーム、多云のクラウド環境、オンプレミスのデータセンター、仮想化されたワークロード、そしてエッジワークロードを含む多様なプラットフォームと環境を持つ。

共通の自動化戦略とは何を意味していますか？

共通の自動化戦略とは、異なる環境すべてを人手を投入するのではなく、自動化によって管理、インストール、運用を実施するというアプローチを意味する。

コンテナ化戦略とは何を行いますか？

コンテナ化戦略は、ワークロードを共通のイメージに収め、異なるプラットフォーム間で非常に類似したワークロードを作成することによって、異なる環境を簡素化する手法を指す。

データの3つの形態とはどのようなものですか？

データの3つの形態とは、静止データ（データベースに格納されているもの）、移動データ（アプリケーション間の通信）、非構造化データまたはコンテンツを指す。

アプリケーション層とは何を意味していますか？

アプリケーション層とは、企業の顧客に対する顔であり、すべての魔法が起こる場所を意味する。カスタムアプリケーションや既存のアプリケーション、第三者アプリケーションが含まれる。

開発者ツール層には何が含まれますか？

開発者ツール層には、CI/CD環境、開発者メトリックス、AIコードアシスタントなどのツールが含まれており、効果的に高い分散環境でチームが働くことを支援する。

AIはどのようにして開発者と協力関係を築いていますか？

AIはコードを書くための素晴らしい助手であり、開発者はまだ具体的な仕様を理解し、テストやアーキテクチャの結果を記述する必要がある。つまり、AIは開発者の仕事全体を担うのではなく、協力関係を築いている。

セキュリティはこのアプローチ全体にどのように関与していますか？

セキュリティは常に私たちの頭の上にあり、物理的なセキュリティからオペレーティングシステム、コンテナ、データ、アプリケーションレベルのセキュリティまで、すべてのレイヤーにわたって考慮される必要がある。

新しい技術にどのように対応する戦略を立てていますか？

新しい技術に対応する戦略を立てることで、現在だけでなく、今後も技術の進化に備えた環境を作り、歴史を理解し、それを基に未来を見据えることが重要である。

Outlines

00:00

😀 IT環境の簡素化物語

ナダンとカイル・ブラウンが、ITの複雑さを簡素化する物語を語ります。ナダンはRed Hatのグローバルチーフアーキテクトリーダーで、カイルはIBMのCTOでもあります。彼らはIT技術の現状と、それを簡素化するプロセスについて話します。IBMの経験から学んだことや、顧客との会話を通じて見られる多くのプラットフォームと環境についても触れています。

05:03

🤖 自動化戦略とコンテナ化の重要性

自動化戦略の重要性を強調し、環境を管理するためには人手を増やすのではなく、自動化が不可欠だと示します。また、コンテナ化戦略についても語り、ワークロードの類似性を見つけ、異なるプラットフォームでも動作するコンテナ化された環境を作り出す方法を紹介します。オープンソースの進化と仮想化戦略への影響についても触れています。

10:08

📚 データの3つの形とアプリケーションの重要性

データは組織を動かす燃料であり、データの3つの形態（静止データ、移動データ、非構造化データ）について解説します。静止データと移動データの管理方法、非構造化データのAIへの活用についても触れています。アプリケーションがエンタープライズと顧客を結ぶ重要な役割を果たし、アプリケーションの簡素化とモダン化の重要性を強調します。

15:12

🛠️ 開発者ツールとAIの役割

開発者ツールの重要性と、CI/CD環境の標準化について話します。マイクロサービスアーキテクチャの普及によりCI/CDプラットフォームの数が増えたこととそれに伴う課題についても説明します。また、開発者にとっての価値創造とAIの役割についても触れ、AIがコードを書くことと開発者が必要なスキルについても議論しています。

🔒 セキュリティとイノベーションへの道

セキュリティが常に重要なトピックであることと、各レイヤーにわたってセキュリティを考慮する必要性を強調します。そして、新技術への対応とその歴史の重要性についても話します。学びとイノベーションの未来に向けた期待を示し、視聴者がチャンネルを購読し、ビデオを高評価にすることを呼びかけています。

Keywords

💡簡素化

簡素化とは、複雑なIT環境をシンプルにすることです。ビデオのテーマは、ITの複雑さを解消し、効率化することに焦点を当てています。例えば、Kyle BrownはIBMが経験した旅について話しており、様々な環境を管理する上で簡素化が重要な役割を果たしていると述べています。

💡自動化戦略

自動化戦略とは、人力を投入するのではなく、システムのインストール、運用を自動化することで環境を管理する戦略です。ビデオでは、Kyleが異種の環境を管理するために自動化が不可欠であると強調しており、これが問題を解決し、さらにはレイヤーの上に何も構築する上で重要な役割を果たしています。

💡コンテナ化戦略

コンテナ化戦略とは、多くのワークロードを少数のパターンに収め、共通のイメージを作り共通のアプローチを取ることです。ビデオでは、IBMがコンテナ化を通じて異なるプラットフォームにまたがるワークロードを非常に類似したものに変え、環境間の移動を容易にしています。

💡仮想化

仮想化とは、物理的なコンピュータを複数の仮想マシンに分割し、それぞれの仮想マシンが独立した機能を持つ技術です。ビデオでは、Kyleが仮想化戦略からコンテナ化にシフトし、コストの問題もその理由として挙げています。

💡データ

データとは、組織を動かす燃料となる情報のことを指します。ビデオでは、データは休止中（データベースに保存されているもの）、移動中（アプリケーション間の通信）、非構造化データ（AIに重要なコンテンツ）の3つの形態で考えられています。

💡アプリケーション

アプリケーションとは、エンタープライズと顧客を結ぶインタフェースであり、ビデオの中心的なトピックです。KyleとNadhanは、アプリケーションの簡素化とモダン化が、開発者にとっては重要であり、顧客のニーズに迅速に対応する上で不可欠であると語っています。

💡開発者ツール

開発者ツールとは、開発者がソフトウェア開発プロセスを支援するためのツールのことを指します。ビデオでは、CI/CD環境の標準化や開発者メトリクス、AIコードアシスタントなど、開発者ツールがどのように開発者の効率を高めるかが議論されています。

💡AI

AIとは、人工知能の略で、ビデオではアプリケーションに組み込まれている最新の技術として紹介されています。AIはデータへのアクセスやトレーニング、ファインチューニングなど、多岐にわたるプロセスに関与し、ビジネスに有用なタスクを実行する能力を発揮します。

💡セキュリティ

セキュリティとは、システムやデータに対する脅威から保護することです。ビデオでは、セキュリティが全てのレイヤーにわたって重要であり、物理的セキュリティからアプリケーションレベルのセキュリティまで、様々な形で議論されています。

💡次世代のイノベーター

次世代のイノベーターとは、academia（学界）や学生など、将来の労働力を指します。ビデオでは、彼らが現在の技術を学び、将来の技術に備える必要があると強調されています。歴史を学ぶことで、彼らはどこへ向かっているかを理解することができます。

Highlights

Nadhan, the global Chief architect leader at Red Hat, and Kyle Brown, an IBM fellow and CTO for the IBM CIO office, discuss simplifying the complexity of IT landscapes.

IBM's journey reflects a typical industry scenario with diverse platforms including IBM Z, cloud, on-prem, virtualized, and edge workloads.

The importance of a common automation strategy to manage disparate IT environments effectively.

Adopting a common containerization strategy to streamline workloads across various platforms.

The influence of containerization on virtualization strategies and the shift towards a more integrated approach.

Managing data in three forms: at rest, in motion, and unstructured data or content, crucial for AI.

The significance of applications as the face of the enterprise to the customer and the focus of simplification efforts.

Modernizing applications to take advantage of containerized and automated environments.

The role of simplification in standardization, easier management, and realizing business value.

The benefits of automation for operators to focus on tasks that require human intelligence and intervention.

The importance of developer tools and managing the explosion of CI/CD environments in distributed architectures.

Creating an environment that supports developers as internal customers to produce more rapidly and relevantly.

Incorporating AI into applications and the need for a strategy to deal with AI, including access to data and application integration.

AI as a helper in writing code, not a replacement for developers, emphasizing the partnership between AI and developers.

Security as a cross-cutting concern that must be considered at every layer of the IT stack.

Addressing the next generation of innovators and the importance of understanding both current technology and historical context for future innovation.

Transcripts

Hello there everyone.

My name is Nadhan.

I'm the global Chief architect leader

in the field CTO organization at, take a guess, Red Hat.

Today we are going to tell a story about simplification,

simplifying the complexity of the IT landscape.

With me here today, I have the privilege

to stand right next to Kyle Brown from IBM.

Kyle.

Thank you very muc Nadhan, I'm Kyle Brown,

I'm an IBM fellow and I'm the CTO for the IBM CIO office.

So the story here is really about, you know, how...

where IT is today, where the technology landscape is today

and we're going to build it up.

You're seeing the the boxes and lines,

but we're going to, as we tell the story,

we're going to write the chapters, write the text.

And at the end of the day, voila,

you're going to have a book on how IT can be simplified.

So, Kyle, when you look at the journey

that IBM has gone through, is going through,

what are some of the, you know, the the platforms,

the environments that you know, IBM has?

And typically that is very typical, I would say, of what we see in the industry.

We are typical.

We're a big company just like most other big companies.

And so if you think about the problem that we have,

we have, for instance, workloads on Z.

Z is at the core of our business

and we do a lot of work on IBM Z platforms.

But Z is not where all of our workloads sit.

We have cloud workloads

not just in the IBM cloud, but in multiple clouds.

So we have to deal with the fact that we have those all over.

We have on-prem workloads

in our data centers that we also have to think about

as we're looking at all of the different possibilities

of the different things that we're running there.

What's more, we also have workloads that are virtualized

that run in all of these different places

on Z, on-prem, on cloud, everywhere.

And then finally, we do have some edge workloads,

particularly in our content management.

We have devices at the edge that we have to control, too.

And if you put this all together,

that basically represents the problem that we face at the CIO.

We have to live with all of these different workloads

and managing all of these different platforms.

Here is what is fascinating, Kyle.

I know this is the IBM story,

but then as you and I, when we go out to meet with customers,

you know, to me this seems very, very typical

of what customers tell they have in their enterprise.

You have in your enterprise.

Absolutely.

As I've talked to a number of customers

and we've talked about these very same set of platforms.

They all agree - that looks exactly like what we have.

And so that begins a conversation that we can start to have

around what did we learn about how to live in an environment

that's distributed across many different platforms like this?

So as we write this book, Kyle, I'm going to take a twist,

you know, a detour, if you will.

Right?

And that is how about if we talk about

how IBM and RedHat come together

with the solutions we have, the approaches we have,

the technologies, on how we can work,

not only just between ourselves, but with our partners

to make lives better for our customers.

Why don't we tell that story?

Sounds great.

So.

Kyle, when you have this many environments

and these are, you know, they come across as singular boxes,

but really, you know, many, many, many instances of these

I cannot even begin to imagine.

How do you deal with that?

What is this thing?

You know that I like the box here,

but what goes in there Kyle?

That was the first of the major decisions

we ended up having to make,

to really understand how to manage something this disparate.

And that is we had to have a common automation strategy.

We realized that you can't just manage all of these environments

by throwing people at the problem.

You have to be able to automate the management,

the installation, the operation

of all of those different pieces of your puzzle.

And so common automation is absolutely critical

to what you need to be able to solve this problem,

and importantly,

to build up any of the layers that come on top of that.

That's great.

Automation is, you know, we will later, we will get into AI,

the two letter magic word.

I would submit, Kyle,

that it's important to have an automation strategy

before thinking about an AI strategy.

So I'm delighted, you know, so everyone watching,

bear in mind if you are talking about an AI strategy,

take a pause.

Ask yourself, how is your automation strategy?

How is our automation strategy?

So great start Kyle. OK, fine,

it's automated all over the place.

What comes next?

So the next decision we had to make

after deciding on a common automation strategy was

we had to come up with a common containerization strategy

because in fact, what we found is that

most of the workloads we have

were all fitting into a very small number of patterns,

meaning that there was an awful lot of similarity

between hundreds of these different workloads.

And what we found is that by putting them into

common images and building a common container approach,

we were able to, in the end, move from having

workloads that were completely different

across these different platforms

into workloads that were very similar.

And the great news is containerization works

across all of these different environments.

Now, you irk my curiosity here.

So did the containerization strategy in recent times

influence the virtualization strategy, by any chance?

It absolutely did for us,

and particularly what we're looking at now are some of the newest changes in open source.

And we are now moving away from a strategy that,

let's say, was a pure virtualization strategy

to one that is instead more tightly integrated with the

backplane of the containerization strategy.

Not to mention that the increasing costs of virtualization

have also moved us to increase our amount of containerization.

We have moved even more strongly

into a fully containerized environment.

So those of you who are watching the video who are dealing with any challenges

related to virtualization, bookmark this

so that you can come back and revisit that for your enterprise.

Great, Kyle, so we have the operating environments,

we have automated, we have containerized.

There is some critical element missing here.

Data.

Absolutely.

You can't have a CIO office without data

because everything runs on data.

It's the fuel that runs the entire organization.

Now when I think about data,

I have to think about it in three different forms.

First of all, there's data at rest.

Now, what that means is that's essentially everything that's sitting in your databases.

You have to be able to not only think about where you want to store everything,

but how you want to manage it, what the standards are around it, and how you deal with it.

Then obviously you've already started working on data in motion.

What are the strategies you're going to have

around being able to make sure that you understand how applications communicate with each other?

What are the kinds of approaches you're going to take b

e they queue driven or event driven or API driven?

You have to work out all of those different pieces.

It's the last one that's a little bit unusual.

When we're talking about data at rest,

and even when we're talking about data in motion,

we're usually talking about structured data.

But that's not all of the data we have.

We have lots and lots of unstructured data or content.

And where that has really become important in the last few years

is this is again the fuel that drives AI.

Being able to have a content strategy for your data,

especially one that involves vectorization

and the ability to take advantage of vector databases and multiple different types of search,

allows you to be able to effectively take advantage of the new capabilities

of new AI models, like large language models.

That's great.

So now when I start thinking about the consumers - the customers, right?

The app, you know, we all have our phones and there are different interfaces.

The application is the face of the enterprise to the customer.

So why are we doing all this?

Yes, it is to simplify.

But to simplify for whom?

It's the end consumer, right?

So I would submit let me take a guess.

This is all about the applications.

And all together I would say especially the containerization is really the platform for application.

Would you agree Kyle?

It is.

And if this is the application layer, and that's where all of the magic happens, is inside of your applications.

the rest of what we've seen beneath that is how we support that layer of applications.

Now that can be both custom applications or off the shelf applications, third party applications.

They still need that support of the underlying containerization and automation layer.

And they need the support of that data layer to be able to do the work that they need to do.

And so what we've done as part of this is we have, as we've talked about, been on this journey

to not only containerize our applications, but to modernize our applications,

to be able to take advantage of this containerized, automated environment.

Excellent.

So, you know, we don't know who is going to be watching the video.

There could be several different roles, right?

So you could be a CIO, or you could be a decision maker and influencer.

So what's in it for you?

What you're seeing here is yes, the you know, the hard truth is that

the technology landscape is complex.

But the approach, the solution and how we address the challenges does not have to be.

You can simplify it.

That's what we heard from Kyle,

how this can and should be simplified.

Simplification leads to standardization.

It leads to, you know, easier management.

You know, and then the KPIs that you have time to deploy realizing value, business value, all of those come along.

The magic word is simplification.

That is what you have, especially if you are a decision maker

in a position to, you know, justify,

rationalize why we have what we have from an IT perspective.

On the other hand, you could be an engineer,

you could be a developer, you could be an operator.

Let's talk about operators.

Life is much easier if the right things are automated

so that you can focus on, you know, what cannot be automated

and what should not be.

Frankly, there needs to be that human touch.

So you can actually focus on the intelligence

of the patterns that you are seeing,

so that when you do root cause analysis, there is actually more meaning to it.

You leverage the automation, you leverage the analytics to do it right.

There are some things that only the human brain can do.

So that's what is in there for operators.

Hello developers.

Let's talk about this here.

And you know, to do all this you need somebody to actually do it, right?

And then write the code and you know, think through the logic

and, you know, the algorithms and all of that.

What is this layer, Kyle?

That's the developer tools layer.

And that's absolutely critical to be able to take advantage of any of the capabilities we've talked about beneath it.

Now, for us, what we found is that there were several different pieces of this developer tool that were really important.

First of all, we had to be able to essentially get a handle on the explosion that we had of CI/CD environments.

One of the things about architecture that has happened over the last few years is that

very deeply distributed architectures, things like microservices architectures, have become very popular.

Microservices architectures are great,

but they have this kind of interesting side effect

in that they make the number of CI/CD platforms you have multiply like rabbits.

They're everywhere.

And it becomes very important to set some standards

and to especially build tools that can help your teams to be able to

work more effectively in that kind of highly distributed environment.

And so what we ended up doing

is we ended up putting together a common CI/CD layer

supporting a number of other tools, including things like developer metrics

and including other things like and how we're getting into AI code assistants.

All of which are supported by a common set of underlying tools

that give you the ability to use the different pieces of this platform.

Excellent, so I would, you know ...

yes, we have been talking about customers, the external consumer or the external customers.

But frankly, from my perspective,

developers are the internal customers.

So if we make these environments easy to use for the developer

so that a minute of the developer's time realizes value thanks to the underlying platforms and the efficiencies that are built in,

the developers can actually produce more,

rapidly, and be more relevant to the features that the customers are looking for.

That's another reason why, you know, IBM is doing all of this, right?

So that the developers can actually, you know, get value out of the environments they are in.

And we live in a world today, if the developers don't get what they need,

they're going to go west, they're going to go elsewhere

and then do things where they actually have the environments they like.

Would do you agree, Kyle?

I completely agree.

And speaking of our friends, the developers,

the one thing that everyone likes now and that everyone is incorporating

into it seems every application in our entire portfolio is AI,

which is our last bottom bucket that we have here.

Now, the interesting thing about AI

is that it follows a lot of the principles

that we've been talking about through these other layers.

You have to have a strategy for how you're going to deal with AI.

You have to have an environment in which you're going to be running your LLM's and in which you're going to be, especially

gaining access to those piles of data

that are important not just for training and fine tuning,

but also for things like the RAG pattern.

And you need to be able to get to your applications, running here,

through things like the REACT pattern to allow you to be able to,

not just have large language models play interesting tracks with language,

but be able to make them do things that are useful to your business.

And so that's the last of the layers that we've seen to be really important here

is you have to be able to have a common, standardized way of doing things

so that everyone doesn't go off and do things in their own way,

which can create not only interesting problems in being able to manage your portfolio,

but especially ethical and security problems.

Now AI can actually write code too, right?

So how would you distinguish between

what would AI come out with versus what the developer still needs to write?

It's not like the developers don't have to write code anymore.

There is a balance there, right?

Absolutely.

There's there's a deep division that we have here

in that, let's say a lot of people think that AI can write all of your code,

but the only people that think that are the people who actually aren't writing the code.

Instead, what we found is I can be a great helper in writing code.

You still have to have very good specifications for it.

You still have to know what the outcomes are and that you want from it.

In other words, you have to be able to specify what you want from your tests.

You have to be able to specify what you want in terms of being able to describe the architecture of the outcome.

And so that's why what we've seen as part of the developer's tools

and as part of the tools that we're building on top of developers tools,

what is instead a partnership

rather than AI taking over the entire business of being a developer

or being a manager or being an operator as part of any of the pieces of this puzzle?

Excellent. So you're the CTO to the CIO, right, at IBM.

So what does the, you know, what does the CISO have to say about it, Kyle?

Right?

Where does security come in?

I have to say that, security is at top of mind for all of us all the time.

And so it's something that cuts across all of these different layers.

We have to think about our physical security that we have.

We have to think about what it means to be securing our operating systems and our containers.

We have to think about data security.

We have to think about application level security.

All of these different pieces

are something where security enters into the equation, and we have to balance that

as part of the overall approach that we're trying to set up.

That's fantastic - as I look at the, you know, the different personas

who could be watching the video, we talked about the decision makers.

We talked about the engineers, both the developers and the operators.

There is one segment that we haven't spoken about.

Those are the next generation of innovators.

From academia, from the faculty, from the, you know, the students.

So those of you who are, you know, in high school and graduate schools and so on, you are the next generation of the workforce.

There is a lot going on here, guys and gals, for you.

Because what you're seeing here is not just the technology of today.

You are seeing how this is positioning for the emerging technologies that are coming around, and also how we got here.

History is very important in understanding where we are going.

So what you heard in this book that we wrote just in the last few minutes

is where we were starting with the Z,

how we wall to on-prem virtualization to the cloud, going to the edge,

and how we have continued to grow vertically leading up to artificial intelligence.

Thanks for watching.

Before you leave, please remember to click like and subscribe.