Disqus - Friends of dave_graham

Re: Del.icio.us Builds Its Social Network

Richard Hintz — Fri, 28 Apr 2006 21:10:03 -0000

So, what are the basic methods for determining:

<ul>

<li>the 10 most influential people in a tag space</li>

<li>the 10 most authoritative people in a tag space?</li>

</ul>

Presumably they wouldn't have to be the same people.

Re: Twitter is Finally Useful for Something

Richard Hintz — Wed, 16 Apr 2008 14:38:07 -0000

Some decent uses of twitter, besides reporting imminent imprisonment, have been emerging.

For Berkeley fans of Cheeseboard pizza, you can follow cheeseboard and get an announcement each morning what the pizza variant will be that day.

Somewhat less useful, since the "channel" is so chatty, is following JetBlue, if you're a JetBlue fan.

Also, since this is the SFist site, it's interesting that the Shanghai "-ist" site, has a twitter feed, shanghaiist, while SFist doesn't.

And lately I've been complaining about problems with different bits of software and I've been getting questions by email about my issues from the software companies, for example, jajah.

Re: I’m crossing my fingers that Orbitz will respond soon

Richard Hintz — Sun, 30 Nov 2008 19:26:59 -0000

Hi, the implication of United's site http://www.united.com/page/... is that they would deal with it, even though it was ticketed by Orbitz. The United call center says otherwise? Maybe call United and refer to their stated policy?

Re: Twitter fails in Vietnam

Richard Hintz — Wed, 10 Dec 2008 23:26:51 -0000

@Clarious Can you use BitTorrent to get the Linux distros?
--
rjhintz on twitter

Re: Cloud Optimized Storage Solutions: Part 3a – Tiering & Expectations

Richard Hintz — Wed, 14 Jan 2009 14:37:47 -0000

Are you going to talk about how an OLTP transaction gets committed to persistent storage in a cloud setting? I'm thinking of, say, a specific airline seat or a financial transaction that represents an equity buy/sell pair.

I (more or less) understand some of the strategies that you outline here and other places once it's ready for commit and what happens to the persistent data after commitment. It's getting to the commit point in a cloud I'm not clear on. Some variant of 2 phase commit? Something with lower latency that's not blocking? Something that's optimistic and only rolls back if X happens?

In this context I guess cloud is synonymous with federated database. I'm asking the federated database people similar questions, but I've bogged them down with security/compliance questions and they've gone stumm.

Re: Interview: Terry Huval

Richard Hintz — Wed, 08 Apr 2009 15:33:38 -0000

Fine interview. Any idea about:
--for the main ring, what fiber was used (LEAF, SMF-28) and count?
--was existing duct used, especially any telco duct? If so, could LUS lease the duct directly or did they need a CLEC?
--for commercial building entrances, does the LUS fiber generally enter from the electric utility side (assuming its diverse from the telco side). Obviously this would be a selling point.
--is LUS running their own 24/7 network operations center? Did they install and are operating their own electronics or is this outsourced?
Thanks.

Re: Interview: Terry Huval

Richard Hintz — Wed, 08 Apr 2009 15:33:38 -0000

Re: MF EOSD

Richard Hintz — Thu, 13 Aug 2009 09:21:00 -0000

Technologic Life
Interesting article. It would be helpful to know:
--current hardware. Is the hardware really 40 years old or just the big ALC application?
--What's the mainframe operating system?
--Why should the replacement application be able to have a 40 year life?
--what's their current business continuity strategy?

There are a zillion other questions raised by the article...

Re: representing non-linearity in project execution

Richard Hintz — Fri, 09 Oct 2009 12:41:48 -0000

Aren't there two different things here? Project management, which shows when a project ends, a planned date, and process management, which shows how a process works.

Project management visualization shows start and end of a project based on start and end of intermediate tasks, some of which may be dependent on others. There's a plan and an actual, where the actual includes unplanned delays, some of which can be accommodated with slack time. Intermediate tasks, perhaps ending in milestones, can have earliest start, planned start, and latest start.

Process visualization can use workflow diagrams, including flowcharts, to show how each of the intermediate tasks gets executed.

Re: representing non-linearity in project execution

Richard Hintz — Fri, 09 Oct 2009 16:53:12 -0000

This model, which I like for other things, does little to schedule when the date/time when the subcontractor to pour the cement needs to arrive on site, even after a weather delay.

What's the asynch message equate to for this?

Re: The financial fallout was an "Inside Job"

Richard Hintz — Thu, 28 Oct 2010 18:48:17 -0000

I just (10/28/2010) heard an interview on MarketPlace Radio with Glenn Hubbard, featured prominently in Inside Job, as an academic facilitator of the thinking that created the environment that led to our current financial problems.

Perhaps he would be a good person to interview on the implications of the points in Inside Job. I'd like to hear some of his perspective about the issues that were raised. If he can't answer any better than in the film, you might want to consider other commentators....

Re: The financial fallout was an "Inside Job"

Richard Hintz — Thu, 28 Oct 2010 18:48:17 -0000

Re: Searching for Buddha in Afghanistan

Richard Hintz — Wed, 08 Dec 2010 17:15:22 -0000

What are the latitude/longitude of the cliff cavities for the former statues?

Re: Searching for Buddha in Afghanistan | People & Places | Smithsonian

Richard Hintz — Wed, 08 Dec 2010 17:15:22 -0000

What are the latitude/longitude of the cliff cavities for the former statues?

Re: My Incomplete, Subjective List of Enterprise SSD Companies

Richard Hintz — Wed, 15 Jun 2011 14:26:11 -0000

Nice legwork. What's your thinking the factors that distinguish enterprise storage from the other classes? (Which are what, commodity, HPC?)

Re: Who’s reading this blog?

Richard Hintz — Tue, 07 Feb 2012 23:57:26 -0000

Hi, I also read through Google Reader, though not always up to date, as with this posting and my tardy reply. I'm especially interested in mapping sites that are mentioned here on Google Maps using Map Maker.

Thanks for this interesting resource!
--
Rich

Re: Show 176 – Intro to Python & Automation for Network Engineers

Richard Hintz — Thu, 20 Feb 2014 16:34:15 -0000

I haven't used this myself, but apparently Salesforce is using Trigger (trigger dot readthedocs dot org)

Re: Show 176 – Intro to Python & Automation for Network Engineers

Richard Hintz — Thu, 20 Feb 2014 20:34:57 -0000

I ran across Stateless Networks, a start up that also wants to play in network automation.

Re: Show 176 – Intro to Python & Automation for Network Engineers

Richard Hintz — Wed, 04 Jun 2014 15:40:15 -0000

Have you looked at GitHub? A quick search on "cisco python" yielded 49 public repositories. Other results for searches on "juniper python," "puppet cisco," and even a couple for "ansible cisco." Bitbucket is another resource, but a search there shows fewer results.

Re: The Platform Abstraction Advantage of PaaS

Richard Hintz — Tue, 20 Jan 2015 15:31:00 -0000

I may be missing something basic, but unless a company has exactly one app with zero data affinities outside the app, PaaS vs IaaS may not be the right question.

What's the goal? To deploy and maintain apps that advance business goals. If the present development/deployment environment is aligned with a credible PaaS supplier, great. Assuming developers can be productive with the PaaS platform and its pace of change, that's an appropriate choice.

The great unwashed, average and below, enterprise developer shops don't fall into this happy category. I read about all the high functioning organizations that are presenting at conferences about their successes and I'm thinking that there are 100 low functioning, barely keeping up companies for each success.

But it's not as if IaaS is the silver bullet solution for the low performing since there's a cultural mismatch. Generally they are as far from DevOps nirvana as one might imagine.

I hope I'm wrong, but the guiding principle in adoption of PaaS or IaaS may well be Darwinism. Those organizations that can innovate will survive. Those that can't will have a different fate.

The question the CIO should be asking is how do we move our code base forward at all, even on our own iron? And how do we simplify / converge the infrastructure, aligning with emerging practice? Most often these questions aren't addressed.

Re: IBM The World's Largest Cloud? Not Even Close - by Matt Asay

Richard Hintz — Thu, 22 Jan 2015 09:10:38 -0000

IBM's SEC 8K filing for 20 January 2015 says that the actual IBM cloud as a service revenue was $3B. The other $4B was "software, hardware, and services to clients to build private clouds," so perhaps there's some double counting.

It seems unfair to say that AWS adds enough capacity to support IBM's claimed $7B cloud business every day. IBM is claiming $7B in revenue (but see above). The Amazon $7B was for the entire Amazon market cap.

Re: IBM The World's Largest Cloud? Not Even Close - by Matt Asay

Richard Hintz — Thu, 22 Jan 2015 10:24:33 -0000

Are you saying that IBM's PaaS and SaaS numbers aren't being counted? The IBM financial reporting says $3B for cloud as a service, which would seem to include these. It probably includes bare metal remote hosting, too, which would really be stretching elastic IaaS.

Re: IBM The World's Largest Cloud? Not Even Close - by Matt Asay

Richard Hintz — Thu, 22 Jan 2015 11:10:07 -0000

I see your point, but why does revenue matter at all to an IT organization, as opposed to investors?

If you're an enterprise your strategic direction might be to move all your workloads to SaaS, if they could be aligned with what the SaaS provider offers, and the rest to PaaS/IaaS.

Realistically, the preponderance of the workloads are legacy and can't be forklifted to SaaS, so you'd want to know the PaaS/IaaS market leaders for all your legacy apps, since revenue is tied to innovation, feature set, and business resilience. You'd choose SaaS mostly on feature set, secondarily on market presence/revenue.

Matt, in saying "while fair" discussing the legitimacy of Microsoft counting Office 365/SaaS revenue as cloud, may be suggesting this split. I haven't seen much discussion counting Salesforce or similar in the same breath as AWS, Azure, or Google. Maybe it's for this reason.

Re: Worst-Case Distributed Systems Design | Peter Bailis

Richard Hintz — Tue, 03 Feb 2015 19:10:40 -0000

Asides 3b. "Making each node in a coordination-free system two times faster via “scale up” techniques makes the whole cluster two times faster."

"If the conversation is typically “scale out” versus “scale up,” if we’re coordination-free, we get to choose “scale out” while “scaling up.”"

Don't understand this. I'm sure I'm missing something basic.

Re: Worst-Case Distributed Systems Design | Peter Bailis

Richard Hintz — Wed, 04 Feb 2015 08:58:41 -0000

Got it, thanks.

I'm in a continuing discussion with mainframe huggers, whose only solution for their primarily financial transaction workloads is scale up. (There's workload sharing with IBM Sysplex, but it's not quite the same as what we think of as distributed computing, especially in a geo distributed, active-active context for persistent data, as AWS enables.)

They, reasonably, point to the scale out coordination requirement as the fatal flaw in proposing a forklift replacement of legacy mainframe with zillions of worker bee, scale out pizza boxes. There are other issues, but this is a fundamental sticking point. Anyway, I'm always interested in scale up / scale out discussion.