The Datacenter is the Computer

Using containers I can easily ship applications between machines and start to think of my cluster as a single computer. Each machine acts as additional CPU cores with the ability to execute my applications and run an operating system, but the goal is not to interact with the locally installed OS directly. Instead we want to treat the local OS as firmware for the underlying hardware resources.

Now we just need a good scheduler.

The Linux kernel does a wonderful job of scheduling applications on a single host system. Chances are if we run multiple applications on a single system the kernel will attempt to use as many CPU cores as possible to ensure that our various applications run in parallel.

When it comes to a cluster of machines the job of scheduling applications becomes an exercise for the operations team. Today for many organizations scheduling is handled by the fine folks running that team. Yet, unfortunately the use of a human scheduler requires humans to keep track of where applications are running. Sometimes this means using complicated error-prone spreadsheets or a configuration management tool with Puppet. Either way these tools don’t really offer the robust scheduling that is necessary to react to these real time events. This is where Kubernetes fits in.

If you think of the datacenter in this way then Kubernetes would be it’s datacenter operating system.

Kubernetes on MesosTry It Now

The inspiration for this post came from Kelsey Hightower (@kelseyhightower).

Myriad is a framework for scaling YARN clusters on Mesos

Myriad is a mesos framework designed for scaling YARN clusters on Mesos.
Myriad can expand or shrink one or more YARN clusters in response to events as per configured rules and policies.

The name Myriad means, countless or extremely great number. In context of the project, it allows one to expand overall resources managed by Mesos, even when the cluster under mesos management runs other cluster mangaers like YARN.

Myriad allows Mesos and YARN to co-exist and share resources with Mesos as the resource manager for the datacenter. Sharing resources between these two resource allocation systems improves overall cluster utilization and avoids statically partitioning resources amongst two separate clusters/resource managers.

Roadmap

Myriad is a work in progress.

  • Support multiple clusters
  • Custom Executor for managing NodeManager
  • Support multi-tenancy for node-managers
  • Support unique constraint to let only one node-manager run on a slave
  • Configuration store for storing rules and policies for clusters managed by Myriad
  • NodeManager Profiles for each cluster
  • High Availability mode for framework
  • Framework checkpointing
  • Framework re-conciliation

https://github.com/mesos/myriad

Open-Source Service Discovery

The problem seems simple at first: How do clients determine the IP and port for a service that exist on multiple hosts?

When developing and running resource-efficient distributed systems like
Apache Mesos (a cluster manager) that simplifies the complexity of running applications on a shared pool of servers, this is a very important decision to make.

Jason Wilder has looked at a number of general purpose, strongly consistent registries (Zookeeper, Doozer, Etcd) as well as many custom built, eventually consistent ones (SmartStack, Eureka, NSQ, Serf, Spotify’s DNS, SkyDNS).

Many use embedded client libraries (Eureka, NSQ, etc..) and some use separate sidekick processes (SmartStack, Serf).

Interestingly, of the dedicated solutions, all of them have adopted a design that prefers availability over consistency.

Please read this really nice writeup by Jason Wilder to learn more.

http://jasonwilder.com/blog/2014/02/04/service-discovery-in-the-cloud/

PHP Next Generation

The PHP Group has put up a post about the future of PHP. They say, ‘Over the last year, some research into the possibility of introducing JIT compilation capabilities to PHP has been conducted. During this research, the realization was made that in order to achieve optimal performance from PHP, some internal API’s should be changed. This necessitated the birth of the phpng branch, initially authored by Dmitry Stogov, Xinchen Hui, and Nikita Popov. This branch does not include JIT capabilities, but rather seeks to solve those problems that prohibit the current, and any future implementation of a JIT capable executor achieving optimal performance by improving memory usage and cleaning up some core API’s. By making these improvements, the phpng branch gives us a considerable performance gain in real world applications, for example a 20% increase in throughput for WordPress. The door may well now be open for a JIT capable compiler that can perform as we expect, but it’s necessary to say that these changes stand strong on their own, without requiring a JIT capable compiler in the future to validate them.’

Keybase.io

I’ve been trying out keybase.io and you can find me at key­base.io/­preillyme.
I think it might be point­ing a use­ful way for­ward on private-by-default com­mu­ni­ca­tion and, for what it does, it gets a lot of things right.

The prob­lem · We’d like to be con­fi­dent that the mes­sages we send across the net  —  email, chat, SMS, what­ev­er  —  are se­cure. When we say “secure” we mean some com­bi­na­tion of “nobody can read them but the per­son who’s sup­posed to” and “the per­son read­ing them can be sure who sent them.” ¶

In prin­ci­ple, this should be easy be­cause of Public-key cryp­tog­ra­phy, which has been around for a while, is re­li­able enough to pow­er ba­si­cal­ly 100% of the fi­nan­cial trans­ac­tions that cross the internet, and for which there’s ex­cel­lent open-source soft­ware that any­one can use for free.

Get­ting cryp­to in place for mail and oth­er mes­sages has been tough, for a few rea­son­s. First, how do you find some­one else’s key re­li­ably, where by “reliably” I mean not just find it, but be­lieve that it’s re­al­ly theirs?

Se­cond, most mes­sages these days live in the cloud (G­mail, Face­book, Twit­ter, etc) and the cloud own­ers like to have them un­en­crypt­ed to help them to ad­ver­tise bet­ter.
So, they’re prob­a­bly not re­al­ly all that mo­ti­vat­ed to help make mes­sages se­cure.

Now, I know that se­cure email is pos­si­ble, and that https con­nec­tions to Face­book and Google and Hot­mail are help­ful, but right now to­day, most mes­sag­ing isn’t very se­cure.

Key­base · Key­base.io does a few sim­ple things: ¶

  • Keeps a di­rec­to­ry of keys that you can look up by a sim­ple name. Since I’m an ear­ly adopter I got “preillyme”, but in prac­tice your email ad­dress would work fine.
  • Lets you prove that the own­er of a key al­so owns a par­tic­u­lar Twit­ter han­dle and Github ac­coun­t. In prac­tice, since I tend to be­lieve that the peo­ple I know are as­so­ci­at­ed with cer­tain Twit­ter/Github ac­counts, I’m in­clined to be­lieve that the keys re­al­ly be­long to them.
  • Lets you en­crypt mes­sages so they can on­ly be read by one par­tic­u­lar per­son, lets you sign them to prove that they could on­ly have come from you, and the in­verse; de­crypt and signature-check.
  • Does all this in a sim­ple web page that’s easy to use, or in a geek-friendly command-line in­ter­face.

So, the idea is that if there’s a mes­sage you want to send, and you want it to be a se­cret, you vis­it key­base.io, paste your text in, en­crypt it for the per­son you’re send­ing it to, sign it, and then copy­/­paste it in­to an email or chat or Face­book mes­sage or what­ev­er. The per­son at the oth­er end copy­/­pastes it in­to key­base.io and re­vers­es the pro­cess and would you look at that, you’ve just prac­ticed se­cure com­mu­ni­ca­tion!

Yeah, it would be bet­ter if this were al­ready built in­to ev­ery mes­sag­ing pro­gram that everyone us­es, and you got it by press­ing a but­ton; or bet­ter stil­l, if ev­ery­thing were al­ways encrypt­ed.

But in the in­ter­im, while this may be a lit­tle klunky, it’s aw­ful­ly sim­ple and easy to un­der­stand; and it works with any­thing that can be used to send a chunk of text from any­where to any­where. So I’m actually pret­ty im­pressed.

In greater depth · Here are a few more tech­ni­cal rea­sons why I like what I see at Key­base. ¶

  • There’s the abil­i­ty to “track” an­oth­er user, which does all the cryp­to check­ing and signs the re­sult, so in fu­ture you can do a quick check whether anything’s changed. This speeds things up and re­moves a few threat mod­el­s.
  • There’s al­so a command-line clien­t, which should be very com­fort­ing for the para­noid. Per­haps the most wor­ry­ing threat mod­el is that some­one shows up at Keybase’s of­fice and, us­ing ei­ther ma­li­cious tech­nol­o­gy, a Na­tion­al Se­cu­ri­ty Agency let­ter, ar­ranges to com­pro­mise their soft­ware; the first time you type your passphrase in­to that com­pro­mised soft­ware, your se­cu­ri­ty is gone. But if you use the command-line clien­t, the ad­ver­sary has to com­pro­mise your own com­put­er to get at you.
  • The ac­tu­al cryp­tog­ra­phy soft­ware is all GPG and Scryp­t; what Key­base of­fers is pipefit­ting and a di­rec­to­ry and some util­i­ties. So the cryp­to part ought to be believably se­cure.
  • It’s all open-source and there on Github. Very com­fort­ing.
  • There’s al­so a REST API, which at first glance looks very sen­si­ble to me.
  • In prin­ci­ple, once the API is locked down, any­one could im­ple­ment a Keybase-style di­rec­to­ry  —  for ex­am­ple to serve a par­tic­u­lar com­mu­ni­ty of trust  —  and mes­sag­ing tools could be taught how to work with any old in­stance.
  • The peo­ple who built this are the ones who built OkCupid, which suggests that their tech­ni­cal chops may well be up to the task.

A wor­ry · You can al­so store your pri­vate key, en­crypt­ed with your passphrase, in the Key­base di­rec­to­ry. This makes cer­tain things eas­i­er and quick­er, but it makes that one par­tic­u­lar threat mod­el, where a bad per­son com­pro­mis­es the soft­ware, even scari­er, be­cause they have your pri­vate key the first time you type your passphrase in­to the com­pro­mised soft­ware.

Trade-offs · If you delete your stored pri­vate key, it means you have to use the command-line client rather than the web in­ter­face. Which is way less civilian-friendly. This is a very, very in­ter­est­ing trade-off. I’m think­ing Key­base is go­ing to have to pub­lish some­thing about their le­gal and po­lit­i­cal de­fen­sive mea­sures. ¶

If you’re us­ing the command-line key­base tool on OS X, you can store your passphrase in the Mac key­chain, so any com­mands that need your passphrase Just Work. So for peo­ple who are handy with the com­mand line, it’s ac­tu­al­ly more con­ve­nient then the Web for­m, which re­quires you to type in the passphrase, or paste it from your pass­word man­ager.

What Does a Technology Strategist Do?

A technology strategist is responsible for developing, implementing and maintaining strategies as they relate to a company’s information technology structure. This is critical for a business because the use of technology typically reduces costs and results in greater efficiency and profit yield. As a technology strategist works, he must coordinate not only with members of management, but also with the company’s clients.

One of the initial tasks an technology strategist has is to evaluate the company’s current technology status. This may include speaking with managers of the information technology team, as well as physically visiting the company’s various centers to assess what is currently in use. The technology strategist also researches what the company’s competitors are doing with technology in order to determine whether his company is being equally innovative. As part of this process, the technology strategist may request formal inventory, technology and technology-related production reports.

During the research process, a technology strategist also tries to determine what the customers of the company need or want that could be supplied or supported with technology. For instance, the strategist might discover that the company’s clients have a strong desire to track shipments. He could take this information to managers of the technology department to develop an order tracking system clients could access online. The strategist may gather information about the client’s needs and wants through techniques such as interviewing, feedback forms, surveys and reviews of the type and number of items sold.

What a technology strategist is able to do for a company is determined to a great extent by the company’s budget and operational philosophies. For instance, if the company has suffered a profit loss in the previous year, the company might not be able to allocate as many resources to technology and technology strategy development. Thus, the strategist has to get information about the company’s current financial status and why the company plans to allocate funds in a specific way before he can design an acceptable strategy. This information usually comes from managers of the finance department, as well as documents such as the company mission statement or project proposals.

Once a technology strategist has all the data he needs about the company, its clients and the company’s competitors, he uses his knowledge about the company’s products, services, market position and current budget and technology status to brainstorm ideas about how the company could approach technology. Each strategy that is developed has to show the risks, benefits, resources and opportunities available for the company clearly, so the strategist spends time getting his ideas in a presentable form, such as a formal report or presentation.

The next step for a technology strategist is to present his ideas to the managers of the company. During this presentation, the strategist’s job is to make sure the managers understand the pros and cons of each strategy proposed. The strategist ultimately does not decide which strategy to follow, but because his insights have a huge amount of sway in the direction the managers take, the strategist has a powerful position within the company.

After the company has selected an appropriate technology strategy, the technology strategist moves on to the implementation and maintenance phases of strategy development. At this point, the strategist concentrates on purchasing and setting up the company’s technology as discussed. If something doesn’t work, the strategist has to troubleshoot and come up with a solution. He has to liaise with clients and suppliers to make this work and monitor results. The strategist may request reports from various departments within the company to evaluate the effectiveness of the plan.

Technology is constantly evolving, so a major challenge a technology strategist has is that his strategies have to be easily adaptable to future technological advances. This is a conundrum because it is so difficult to predict what technology will arise or to determine by what point any new technology will be truly functional. Subsequently, even after the strategist has developed and implemented a technology strategy, he has to be on the lookout for more efficient options constantly. The strategist plays a key role in helping a company decide when and how to switch from one technology to another if necessary.

http://www.linkedin.com/in/preilly

Get Ready For a Streaming Music Die-Off

Streaming services are ailing. Pandora, the giant of its class and the survivor at 13 years old, is waging an ugly war to pay artists and labels less in order to stay afloat. Spotify, in spite of 6 million paid users and 18 million subscribers who humor some ads in their stream, has yet to turn a profit. Rhapsody axed 15% of its workforce right as Apple’s iTunes Radio hit the scene. On-demand competitor Rdio just opted for layoffs too, in order to move into a ‘scalable business model.’ Did no one wonder about that business-model bit in the beginning? Meanwhile, Turntable.fm, a comparatively tiny competitor with what should have been viral DNA, just pulled the plug on its virtual jam sessions this week—and it just might be the canary in the coal mine.