Latency on SWiK

Real-Time Events

Wed, 04 Jun 2008 22:51:39 -0700

Opher reports a car accident (unplanned events again) and concludes that people need to process events in real-time and not in batch.

Congratulations to Opher on his fast reactions, and commiserations on the slower reactions of the driver behind. Clearly there are some events you have to process in real time. But I hope he is not implying that all events must be processed in real-time. When the fuel tank indicator appears, do you refuel immediately or do you wait until you reach the next gas station? How often do you have the vehicle serviced?

I think the critical question for systems designers here is to determine which are the events that call for a real-time response, and which are the events where a batch response is more appropriate.

There are also events that call for a low-latency (extremely fast) response, but don't count as real-time. For example, in the financial markets, there may be short-lived arbitration opportunities, which means you can make a lot of money if you can react within a small number of milliseconds. This is not a real-time requirement, because nobody expects you to pick up every single opportunity - just catch a reasonable number of them. (The reactions of the frog should allow it to catch just enough insects to fill its belly - but some insects escape to breed more insects. The insects get faster, and so do the reactions of the frogs.)

Surely the event infrastructure should be capable of handling any of these patterns.

Mark to Market

Fri, 03 Oct 2008 02:58:28 -0700

Politics and Economics

There have been a number of calls by politicians around the world for an end to the financial practice of "mark-to-market".

Defenders of the practice try to explain that "mark-to-market" simply means taking latency out of the accounts of individual companies, and improves overall transparency. ["Mark-to-market" accounting fight goes down to the wire]

Opponents of the practice point to its effect in amplifying the recent catastrophic falls in share values, particularly banks. European politicians (including David Cameron and Nicolas Sarkozy) have suggested the practice should be suspended or modified [Sarkozy seeks change to fair value rules, Nicolas Sarkozy's mission impossible?] and the US Senate has now given the SEC powers to suspend mark-to-market.

Financial regulations based on the mark-to-market principle were introduced after Enron. [Enron-era accounting reforms blamed in financial crisis] Some politicians (including Mark Sanford) have pointed out a troubling pattern of yesterday's solutions causing today's problems. [A Bailout for All Our Bad Decisions?]

Implications for SOA

Some SOA champions in the financial sector have seen the elimination of latency as one of the critical benefits of SOA. These benefits might now seem contingent and short-term rather than fundamental and lasting. Solving one set of problems, but contributing to new problems.

There is now a real possibility that regulators may impose new and more complicated valuation rules for complex financial assets. If SOA has been correctly implemented, then plugging new valuation rules into existing applications will be a fairly quick and painless procedure. However, if it turns out that the service-oriented architecture needs radical restructuring to accommodate such changes, then this might indicate that the SOA applications weren't correctly architected in the first place.

SOA was supposed to make business more capable of withstanding change. So are banks with a decent service-oriented architecture now in a (slightly) better position to weather the catastrophe? Perhaps we shall soon find out.

Biographical Notes

David Cameron is the leader of the UK Conservative Party.
Nicolas Sarkozy is the President of France, and former leader of the centre-right UMP party.
Mark Sanford is the Republican Governor of South Carolina.

[from phil1eng] It's the Latency, Stupid

Wed, 04 Jun 2008 23:53:22 -0700

Citrix`s Smooth New XenDesktop

Wed, 21 May 2008 01:24:06 -0700

LatencyTOP

Tue, 22 Jan 2008 06:28:57 -0800

LatencyTOP

Sun, 20 Jan 2008 15:51:33 -0800

Tool for monitor latency in linux

LatencyTOP

Sun, 20 Jan 2008 07:51:23 -0800

LatencyTOP

Sun, 20 Jan 2008 05:51:35 -0800

LatencyTOP

Sat, 19 Jan 2008 20:50:26 -0800

LatencyTOP

Sat, 19 Jan 2008 18:50:30 -0800

LatencyTOP

Sat, 19 Jan 2008 17:51:34 -0800

PCI Latency - MythTV

Sat, 19 Jan 2008 16:27:37 -0800

LatencyTOP

Sat, 19 Jan 2008 15:50:19 -0800

LatencyTOP is a Linux* tool for software developers (both kernel and userspace), aimed at identifying where in the system latency is happening, and what kind of operation/action is causing the latency to happen so that the code can be changed to avoid the

LatencyTOP

Sat, 19 Jan 2008 14:51:03 -0800

LatencyTOP

Fri, 18 Jan 2008 19:50:34 -0800

PCI Latency - MythTV

Fri, 04 Jan 2008 18:54:13 -0800

Bobsguide - INTEL LOW LATENCY TRADING LAB SET TO IMPROVE FINANCIAL TRADING PERFORMANCE

Thu, 15 Nov 2007 08:57:18 -0800

ZFS RAID recommendations: space, performance, and MTTDL : Ramblings from Richard's Ranch

Wed, 03 Oct 2007 03:56:56 -0700

SmokePing net latency charting lib

Thu, 13 Sep 2007 13:29:02 -0700

SmokePing - About SmokePing

Fri, 24 Aug 2007 21:05:08 -0700

QoS - voip-info.org

Thu, 02 Aug 2007 21:27:12 -0700

QoS (Quality of Service) is a major issue in VOIP implementations. The issue is how to guarantee that packet traffic for a voice or other media connection will not be delayed or dropped due interference from other lower priority traffic.

[from amaah] The Challenges of Latency

Sat, 05 May 2007 21:43:31 -0700

In any large-scale system, there are a few inescapable facts: 1. A broad customer base will demand reasonably consistent performance across the globe. 2. Business continuity will demand geographic diversity in your deployments. 3. The speed of light isn

PCI Latency - MythTV

Sun, 25 Mar 2007 03:00:57 -0700

bookmark this on del.icio.us - posted by nhivanye to video howto linux myth mythtv pvr latency - more about this bookmark...

Oracle on Opteron with Linux-The NUMA Angle (Part III). Introducing the Silly Little Benchmark.

Tue, 30 Jan 2007 10:09:08 -0800

In my blog “mini-series” about Oracle on Opteron NUMA, I am about to start covering the Linux 2.6 NUMA API and what it means to Oracle. I will share a lot of statspack information for certain, but first we need to go with micro-benchmark tests. The best micro-benchmark test for analysis of memory latency is [...]

Oracle on Opteron with Linux-The NUMA Angle (Part III)

Mon, 29 Jan 2007 18:06:36 -0800

Optimizing Page Load Time - die.net

Mon, 30 Oct 2006 14:21:51 -0800