FINITE INFINITY (Second Edition

Description

• cours - matière potentielle : lvii
• leçon - matière potentielle : clock
• exposé

1 In the first place, we entirely shun the vague word space, of which, we must honestly acknowledge, we cannot form the slightest conception. Albert Einstein, 1920 Nur die Fülle führt zur Klarheit, Und im Abgrund wohnt die Wahrheit. Friedrich Schiller FINITE INFINITY (Second Edition) By D. Chakalov, 1 February 2012 Online version here

• inversion of points
• archimedean volume of space
• imagine that the stars
• physical universe
• finite infinity
• finite volume of space
• unknown unknown
• grin of the cat without the cat
• cannot
• space

Sujets

• Exposé
• Helfer
• Neisser
• Norris
• Leacock
• Armstrong
• Carroll
• Schiller
• Thompson

Informations

• Publié par: miphug
• Nombre de lectures: 31
• Langue: English

Extrait

Navigating the Road to an Automated Application Cloud

Why GPS-based Navigation Offers Answers for Modern IT Automation

How GPSRevolutionizedRoad Travel

Most of us recall the days before global positioning systems (GPS) became the savior for modern road travel. Now nostalgic reminders of bygone days, yesterday’s roadmaps can be found among Route 66-variety memorabilia— glass Coca-Cola bottles, antique gasoline signs and other artifacts from simpler times.

Of course, roadmaps themselves were far from simple. In fact, they were often mad- deningly complex, inscrutable documents that led to wrong turns, missed deadlines, costly course corrections and more than a few marital disputes.

The first innovation to improve upon roadmaps was the TripTik, a customized book- let of turn-by-turn directions based on a specific origin and destination provided to members of the Automobile Association of America (AAA). It removed much of the complexity of fumbling through maps, but offered no assurances you wouldn’t get lost on the road from point A to point B. When you got off track, the TripTik was rendered useless.

Fast-forward to 2011— and the idea of using a roadmap or a TripTik as the basis for navigation seems as outmoded and unnecessary as thermal fax paper.

Today, you simply add a destination to a GPS device and it does the hard work.

Getting from 321 Peachtree Avenue in Atlanta to 123 Main Street in Chicago may require hundreds of individual instructions (and, thus, hundreds of opportunities for human error), but it now becomes a simple command to a GPS device, which is then counted upon to easily guide you to the Windy City. If you take a wrong turn along the way, the GPS puts you back on track – automatically.

The approach is called “declarative” automation, which is distinguished from other forms of automation by shifting the focus from the detailed steps (the “how”) to the desired outcome itself (the “what”). A declarative approach focuses on the end-state, relying on intelligent robots to carry out the fulfillment duties.

How do the robots know how to get you there? They use detailed models and location- based satellite systems that allow you to triangulate on the destination, irrespective of the starting point or any intermediate point along the way.

GPS has created—to use a technologist’s term—an “abstraction layer.” As the driver, you’re spared from the pain and allowed to focus on more important tasks, like keeping your eyes on the road.

Copyright © 2011 rPath, Inc.

if you take a wrong turn along the way, the GPS puts you back on track – automatically.

2

WhatGPSMeans for IT Automation

This GPS example illustrates how automation can be applied to improve the outcome of complex, dynamic situations like long-distance road travel. Because of GPS, we get to our destinations faster and more predictably. It’s a major step forward in the state of the art of road transportation.

Today, most IT organizations are still navigating by map, but the road conditions are more hazardous and the posted speed limits much, much faster.

How applications and other software systems are built, configured, deployed and changed is something like a pre-GPS road trip: IT personnel is counted upon to execute hundreds of discrete tasks to move an application from development into production—hundreds of steps, hundreds of opportunities for error.

It’s a harrowing process that can takes weeks or months as a system moves across specialized silos—OS, middleware, application and security teams all have their hands in the process. In the meantime, business lines wait. And wait.

The Amazon Effect

That a bookseller in Seattle revolutionized how IT is delivered may be among the most unlikely and universally accepted truths in modern business. As we’re all now well aware, Amazon Web Services allows developers to deploy infrastructure in minutes, which is something that takes weeks in most enterprises.

This remarkable fact shines a light directly on enterprise IT and traditional managed service providers, creating new expectations for responsiveness and forcing IT providers to automate more deeply and completely than ever before.

In the same way that GPS became absolutely vital to the competitiveness (and survival) of freight carriers and logistics providers, the same depth of automation is becoming absolutely essential to in-house and outsourced IT providers.

Toward an Application-Centric Data Center

For IT, Amazon has demonstrated an application-centric operating model—one that abstracts and automates much of the complexity of underlying infrastructure and enabling software so developers can focus on the one and only thing that matters to business: applications.

Of course, today, many traditional IT providers are still mired in the muck.

Copyright © 2011 rPath, Inc.

IT personnel is counted upon to execute hundreds of discrete tasks to move an application from development into production— hundreds of steps, hundreds of opportunities for error.

3

While virtualization and infrastructure-as-a-service (IaaS) automation seeks to reduce much of the time and human labor in standing up infrastructure, applications themselves—what business cares about—are still packaged, configured, deployed and changed by hand or with shallow scripts.

Increasingly, manual effort is giving way to custom and commercial script-based solutions, but IT is finding that scripts—in either form—simply don’t scale.

Why ScriptsFall Down

The rise of virtualization and cloud computing has collided headlong with agile develop- ment techniques, creating a perfect storm of circumstances that have brought IT to the brink. Virtualization and cloud are compounding the scale of systems in production, at the same time that Agile—and a growing dependency on IT-enabled business models—are increasing the pace of new applications and application changes released into production.

While most IT providers now acknowledge the need to automate, most have responded with script-based solutions. If roadmaps are the bounded standard operating procedures for the manual data center, scripts are the TripTik.

Scripts are programmatic instructions written by talented and motivated IT professionals to help automate manual processes. Scripts are nothing more than small software programs that are designed to automatically execute the specific steps and procedures that were previously performed by hand.

The goals behind these efforts are certainly correct, but the reality is that scripts are brittle, complex and error-prone. And in the age of declarative approaches to auto - mation—the enablers of the GPS model—they’re completely unnecessary.

Garbage In, Garbage Out

There’s an old adage that automation can make the wrong things happen, faster.

Scripts aren’t intelligent—they’re precise instructions that can’t adapt to changing circumstances. Like a TripTik, scripts won’t help you if you make a wrong turn. Once you’ve departed from the garden path, you’re on your own among the tumbleweeds.

As any other software program, scripts are complex, brittle and opaque. When the instructions are incorrect, they’ll inevitably take you to the wrong destination.

Scripts get out-of-date and buggy and new ones need to be written to accommodate most any change in process or circumstance. That means, as complexity grows, so

Copyright © 2011 rPath, Inc.

4

do the number of scripts.

Ironically, after writing these scripts to manage software, many IT providers face a new management burden: managing the scripts themselves!

In the age of speed, scale and change, a script-based approach to automation just can’t keep pace with the needs of IT. Instead, IT needs what is the equivalent to a GPS device—an intelligent and declarative tool for automating applications and tak- ing the time and risk out of the journey into production.

A GPS forApplication Automation

The key enablers behind a GPS are declarative models, which are like blueprints that map the contours of the terrain; and satellite-based triangulation (or, technically, trilateration, which is a variant of the same concept that measures distance instead of angles). Models provide the lay of the land and triangulation determines precisely where you are at any moment in time.

The combination of these two ideas is perfectly suited for automating complex tasks in dynamic circumstances. It provides technical correctness, contextual awareness and the ability to adapt to changing circumstances.

It makes automation, in a word,intelligent.

At first blush, the GPS analogy may appear to be a modest stroke of poetic license— a device stretched to its limit in the service of storytelling. After all, software sys - tems aren’t like airplanes, ships and automobiles, which cross boundaries of time and space. Right? Actually, while software systems don’t share all attributes of their atom-based counterparts, they do move…

• Over time—as operating systems, middleware, applications and configurations evolve based on their own independent lifecycles;

•Through space—as applications move across public and private clouds, data centers and hosting facilities.

That’s why GPS is, if not perfect, a very relevant analogy for what IT requires.

Copyright © 2011 rPath, Inc.

5

Applying theConcepts: Lessonsfrom Modern Manufacturing

If GPS is the basic inspiration for a more intelligent approach to application automation, modern manufacturing practices offer the architectural direction.

The rise of global competition has forced today’s manufacturers to deeply automate to improve market responsiveness, drive down costs and improve quality. Of course, similar factors have forced IT organizations to automate.

Despite our tendency to think of software delivery as perhaps more art than science, the reality is that it is really a specialized form of manufacturing.

Like manufacturers, IT providers…

• Manage a supply chain(custom, commercial and open source software); • Execute a construction process(dev, test, deployment and change); • Are accountable to quality standards(security, performance, availability).

Likewise, manufacturers have used automation to achieve innovations that have the potential to be equally transformational for IT providers. Enterprise IT and managed service providers should look to the manufacturing example as the basis for automating the application delivery process. Consider the following:

Speed and quality

Like manufacturers, IT providers should embrace deep automation of the supply chain and the construction process to improve throughput, responsiveness and to eliminate the variability and quality issues associated with human involvement.

Just-in-time manufacturing

Manufacturers dramatically minimize inventory holding costs by shifting the focus from finished goods to virtual descriptions—blueprints, bills of materials, etc. This allows new products to be “stamped out” on a just-in-time basis, eliminating the cost and risk of holding finished inventory. Similarly, IT providers should shift the focus from managing system images to managing models, detailed blueprints that describe the content, configuration and metadata of a system or application.

Mass customization

Manufacturers standardize foundational product components, but tweak how they’re

Copyright © 2011 rPath, Inc.

6

composed and configured to address specific market requirements. Standardization, together with just-in-time methods, enables mass customization, which allows manufacturers to take advantage of both scale economies and the flexibility to address diverse market requirements.

These same concepts have significant implications for IT providers. Focusing on detailed models (managed under version control) enables IT providers to create custom system definitions to address diverse user requirements without allowing unmanageable image inventories to grow exponentially.

Decoupling of design from manufacturing

If you look closely at Apple’s packaging, you’ll find the following statement:

Building with Legos: The Netflix Approach to Cloud Automation

Designed by Apple in California. Assembled in China. Netflix faces one of the most demanding scale challenges on the planet and they’ve distinguished themselves as Apple is a product design company, not a manufacturing company. They’ve built their a savvy scale operator with a deep brand and competitive advantage on product design and specifications, outsourcing the focus on automation. Greg Orzell from supply chain and the manufacturing process. This is increasingly common for product Netflix engineering blogged about the companies and it offers lessons for today’s IT providers. transformation of the Netflix build and release process into something that closely resembles a manufacturing The iPhone, for example, is a complex device that encompasses hundreds of production line, focusing on the individual components that are assembled, integrated and packaged into a finished architecture of how systems are product. Detailed blueprints, bills of materials and specifications guide these assembled, integrated and configured components from a bag of parts into an elegantly composed system. and the control mechanisms for how these systems are managed over time.

Each of these components has its own lifecycle and will change over time, which According to Orzell, “When we started impacts the state and stability of the shipping product. Guiding the process of migrating our systems to the cloud assembling, integrating and changing these components are detailed models tightly we took the opportunity to revisit our complete build pipeline, looking both specifying how the parts should come together into a functioning whole. at how we could leverage the cloud paradigm as well as the current landscape This decoupling of design and manufacturing also supports multi-sourcing strategies for build tools.” where manufacturers engage with a network of outsourced supply chain and manufacturing partners. At the heart of this process transformation is a focus on model-driven automation which, among other things, Orzell This example should bring to mind the software supply chain driving the IT delivery says, “…guarantees that what we test process: the inputs are custom, commercial and open source components, which in the test environment is the EXACT are assembled, integrated and composed into functioning applications, which are same thing that is deployed in production; there is very little chance of configuration deployed to serve some business need. or other creep/bit rot.” The Netflix lesson? Cloud-scale computing requires deep, Similar to Apple and so many other product manufacturers, IT providers should focus industrialized automation. on well-defined blueprints that describe how applications come together. Like the Read the full post:http://techblog. GPS example, these models drive intelligent automation that ensures you reach your netflix.com/2011/08/building-with- intended destination safely and on time. legos.html 7 Copyright © 2011 rPath, Inc.

Similarly, this focus on blueprints allows IT providers the flexibility to stamp out new applications to run in virtually any deployment environment. As enterprise IT pur - sues its own multi-sourcing strategies, this approach makes it easy to rapidly deploy, change and migrate applications across a constellation of public and private physical and virtual computing environments, providing the application automation to enable a true hybrid cloud strategy.

WhyManagingImages Doesn’t Make Sense

The previous discussion established why manufacturers have shifted away from finished inventories to blueprints and bills of materials: the cost of holding inventory is an unnecessary burden in the age of just-in-time production.

Why, then, do IT providers still manage finished system images? System images, like finished product inventories, are exceptionally burdensome to manage and impossible to maintain in a cloud environment.

A shift to model-driven automation sets IT providers free from images.

Similar to just-in-time production techniques, a model-driven approach allows images to be stamped out on demand from version-controlled blueprints. This means faster image construction, seamless change, lower storage costs and the ability to quickly retarget systems across deployment environments.

Today, system images proliferate as a negative side effect of virtualization. The low effective cost of hardware makes it easy to justify new workloads, taking the economic friction out of provisioning. Additionally, cloning image templates makes it simple to stand up new applications. Of course, the growing diversity of user requirements means that image templates are often cloned and tweaked before they’re deployed, which sets into motion an unmanageable and risky dynamic where images proliferate like unpredictable gremlins in our midst.

So, as the volume of system images explodes, so do the management challenges they pose. Images are brittle, opaque and difficult to change. They’re deployment vehicles, not assets with a manageable lifecycle. Just the same, we warehouse images and try to manage the unmanageable. A better approach is to shift the focus from images to blueprints, much like modern day manufacturers.

Copyright © 2011 rPath, Inc.

a shift to model-driven automation sets IT providers free from images.

8

IntroducingrPath Cloud Engine

rPath Cloud Engine is intelligent, model-driven automation for rapid construction, deployment and change of software systems. rPath generates images on demand for rapid deployment and retargeting of applications. And by shifting the focus from scripts and images to declarative models, rPath takes the time, complexity and risk out of application deployment.

Key rPath capabilities:

• End-to-end application automation—rPath automates images and application delivery, which allows you get apps and updates deployed correctly and fast.

• Model-driven approach—rPath creates a detailed blueprint as the basis for application deployment and change over time. This virtually eliminates the risk of deployment failures and change conflicts related to software configurations.

• Version-control—rPath provides a deep and detailed version history of your applications that provides control, reproducibility and compliance over time.

• Multi-tenancy—rPath enables multi-tenant environments by allowing  application stacks to be segregated as independent version-controlled instances.

• Plug and play—An open architecture helps you seamlessly integrate rPath into any cloud technology ecosystem to deliver differentiated value.

Copyright © 2011 rPath, Inc.

9

Conclusion: Toward anAutomated Application Cloud

As speed, scale and the rate of change collide in today’s data center, manual methods and script-based approaches (yesterday’s maps and TripTiks) must give way to more intelligent, automated approaches.

Of course dozens of IT management vendors claim to have the automation you need. But beware of wolves in sheep’s clothing! Many commercial automation products including runbook automation, orchestration tools and configuration engines

are often little more than a slick UI in front of a pile of scripts.

These script-based approaches focus on the “how” of application deployment. Like a GPS device, rPath allows you to focus on the “what”:

•“Boot a new application server” •“Provision the brokerage application” •“Upgrade the application to the latest available version” •Roll back the application to the previous version

Like a GPS device, rPath makes it as simple as asking the question and it ensures you never get lost along the way.

At the heart of this approach are the lessons learned by the manufacturing efficiency experts that taught us long ago that deep automation is essential for survival in the age of massive scale, complexity and change.

Today’s IT providers are wise to take these lessons to heart.

###

like a GPS de- vice, rPath makes it as simple as asking the question and it ensures you neverget lost along the way.

10

rPath is the leading private cloud technology provider for rapid application delivery. Named a “Cool Vendor” in release management by Gartner in April 2011, the rPath Cloud Engine is built on the battle-tested system automation platform of rPath x6 and features end-to-end application management. The solution boasts a model-driven approach that ensures rapid application delivery and eliminates vendor lock-in. rPath Cloud Engine seamlessly integrates into any cloud technology ecosystem, helping cloud providers, builders and buyers ensure a cloud that delivers consistency, reliability and speed. Headquartered in Raleigh, N.C., rPath customers include many of the world’s largest enterprises and managed service providers.

For more information, please visitwww.rpath.com.

Copyright © 2011 rPath, Inc. All rights reserved. The information contained in this document represents the current view of rPath on the issue discussed as of the date of publication. Because rPath must respond to changing market conditions, it should not be interpreted to be a commitment on the part of rPath, and rPath cannot guarantee the accuracy of any information presented after the date of publication.

This white paper is for information purposes only. rPATH MAKES NO WARRANTIES, EXPRESS OR IMPLIED, IN THIS DOCUMENT. rPath may have patents, patent applications, trademark, copyright or other intellectual property rights covering the subject matter of this document. Except as expressly provided in any written license agreement from Macromedia, the furnishing of this document does not give you any license to these patents, trademarks, copyrights or other intellectual property.

rPath, rBuilder, rPath Appliance Platform and Conary are trademarks trademarks of rPath, Inc. in the United States and/or other countries. The names of actual companies and products mentioned herein may be the trademarks of their respective owners.

rPath, Inc. 5430 Wade Park Boulevard, Suite 310 Raleigh, NC 27607 866–508–6200

11