Starting now, we will post all future content only at Perspectives. Please update your bookmarks accordingly. To update your RSS feed reader, click here.

Of course, compared to this messaging-specific blog, there is much wider array of content at Perspectives, across all the Informatica product lines, so we encourage you to explore a little bit.

Thanks for reading, and we hope you will migrate over to the Informatica Perspectives blog along with us!

This post will stay at the top of this blog indefinitely. As for the archived content here, we expect to migrate that into the Perspectives blog in the near future.

Also, if you use Twitter or Facebook, and would like to follow/friend us, you can do so by clicking here (Informatica Twitter page) or here (Informatica Facebook page). FYI, all Perspectives blog posts are published in both places, so this can be a convenient way to be notified of blog updates, along with all the other information you can see there. Personally, I find both Twitter and Facebook to be quite useful for this feature, and this seems to be the trend for the future, as well.

Other links:

• Informatica YouTube channel
• Informatica Flickr page
• Informatica Twitter RSS feed

You can reduce disk latency by using only the outer portion of your hard drive, with the obvious tradeoff  of less available disk space. This is called “short stroking”, and is described in some detail in this article “Accelerate Your Hard Drive By Short Stroking”.

Networking folks might enjoy packetlife.net, a site whose goal is to “offer free, quality technical education to networkers all over the world, regardless of skill level or background”.  The site has lots of content — including a forum, a blog, and much more — and is probably worth a bookmark.

If you are interested in memory management or buses, this article might be worth your time: NUMA-aware heap memory manager by Patryk Kaminski of AMD.

Links:

• Accelerate Your Hard Drive By Short Stroking at www.tomshardware.com
• packetlife.net
• NUMA-aware heap memory manager at developer.amd.com

NEW ARTICLES:

• Ultra Messaging Integration with Microsoft^® Excel^®
• Wildcard topic regular expressions and Rendezvous wildcards
• How to create your Support account
• LBM 4.0 change log
• rv2regEx.c sample code added to KB Downloads section — demonstrates a method of converting TIBCO^® Rendezvous^® wildcards into regular expressions, suitable for Ultra Messaging applications
• ss2re.c sample code added to KB Downloads section — demonstrates a method of converting TIBCO^® SmartSockets^® wildcards into regular expressions, suitable for Ultra Messaging applications

UPDATED ARTICLES:

• Wildcard topic regular expressions and SmartSockets wildcards now points to new sample code file (see above) ss2re.c – please note, this article was formerly titled “Wildcard Topic Regular Expressions”
• Address and Port Usage updated for style and clarity (not content)
• LBM 4.0 compatibility and migration checklist now includes new information re: sending from context thread
• Zero Object Delivery (ZOD) updated to fix minor errors

For those who don’t know, SIFMA is an annual show put on by the Securities Industry and Financial Markets Association, an organization which “brings together the shared interests of hundreds of securities firms, banks and asset managers”. High on that list of shared interests, of course, is ultra-low-latency messaging technology, which is precisely what 29West-Informatica’s Ultra Messaging is all about.

To all the prospects and customers who stopped by our booth, thanks!  To those that didn’t, hope to see you next time!

Pictured at the right is our booth before the opening of the show, featuring Matt Saxe, Sales Director for the Northeast region, perfecting his delivery for the camera. In the background is Systems Engineer Gideon Goldman, doing actual work, as usual.  Or, maybe he’s checking World Cup scores. It’s pretty hard to tell.

Links:

• LBM 4.0 GA Announcement, plus ULB and UMQ –  LBM 4.0 blog post
• SIFMA About page

Press:

• 29West Tests LBM Update
• SIFMA 2010: Market Data Must-Sees
• Informatica Releases New Ultra Messaging Offering; Latency Busters Messaging 4.0 Delivers Breakthrough Performance for Financial Services Operations (Globe Newswire).  Same release at TheStreet.com, Marketwatch, Financial Tech Spotlight, and CNN Money.

First is LBM 4.0, featuring big performance improvements with the IPC transport, for applications on the same host.  These performance gains are highly scalable, as well.  As a result, with LBM 4.0, you can now run a complete trading system on a single server and eliminate excess latency from the system.

In addition to LBM 4.0, 29West-Informatica also announced these two other Ultra Messaging offerings, also generally available:

• Ultra Low-Latency Load Balancing (ULB) – Fully leverage parallel processing for demanding front-office messaging applications, such as high-frequency trading, with the nearly-horizontal scalability across multiple CPUs or servers provided by Ultra Messaging.
• Ultra Messaging Queuing Edition (UMQ) – A brand-new product, UMQ extends low-latency, once-and-only-once messaging to back- and middle-office applications, along with the performance and reliability advantages of Ultra Messaging. UMQ enables more flexibility in application design with fault-tolerant queuing, low-latency load balancing, and multiple dissemination models.

See links below for more, or contact your 29West-Informatica sales representative.

Links:

• LBM 4.0 Features Include Latency as Low as 660 Nanoseconds! –  LBM 4.0 blog post
• IPC Performance Report Shows Ultra-Low Latencies of 660 Nanoseconds –  IPC performance report blog post
• Informatica Releases New Ultra Messaging Offering; Latency Busters Messaging 4.0 Delivers Breakthrough Performance for Financial Services Operations –  Press release

The good news is, this particular performance problem can be fixed in a simple way: by using double-buffering, which avoids unnecessary delays caused by said interaction, as explained below. Quoting from the article:

Send your first request, and while you’re still waiting for the response, generate and send your second. Then when you get the response for the first, generate and send your third request. This way you always have two requests outstanding. While you’re waiting for the response for request n, request n+1 is behind it in the return pipeline, conceptually pushing the data along.

But read the whole article, it’s well worth your time, and isn’t all that long. Plus: nice graphs, and who doesn’t like a nice graph? Nobody.

Links:

• “TCP Performance Problems Caused by Interaction between Nagle’s Algorithm and Delayed ACK” by Stuart Cheshire

29West-Informatica ZLT London Event (June 2010)

Many thanks to all who registered for our ZLT London event, held on June 8.  Sure, the vicious rainstorm and uncooperative public transit closures didn’t help us out much, but even through all that, over 70 attendees made it to the venue at Nomura House.

Click the picture to see a full-size image from the event.  Click here to see the wrap-up page on our website.

Don’t forget to check the ZLT page on the website and this blog for future ZLT events, such as Chicago in September and New York in October/November.

A-Team Citihub Video Series: Low Latency Connectivity Solutions

Jitesh Ghai, Global Director of Systems Engineering for 29West-Informatica, explains the strategic importance of low-latency messaging middleware within the high-performance trading environment, in this video interview conducted by A-Team Group.

This version is a major release, and includes several important new features, including:

• 75% improvement in latency - as low as 660 nanoseconds with the LBT-IPC transport (for processes on the same host)
• Topic Resolution improvements – more flexibility and lower overall network bandwidth utilization
• Myricom Datagram Bypass Layer (DBL) support - in conjunction with DBL-enabled 10GigE NICs for Linux and Microsoft^® Windows^®

In addition to these features in LBM 4.0, 29West recently announced the integration of the Voltaire Messaging Accelerator™ (VMA) into our Ethernet transports to allow for latencies significantly under 10 microseconds, to go along with earlier support for Infiniband “native” verbs. No code changes are necessary for 29West Ultra Messaging^® applications. Customers now have a choice of data center fabrics that deliver low latency for their high performance messaging systems.

Contact 29West-Informatica for more information on this version, or click here to download if you hold a current LBM license.

In the future, Latency Busters^® Messaging (LBM) will be known as Ultra Messaging^® Streaming Edition.

The guide is intended to do two things: (1) show the tradeoffs that you should consider when you prioritize lowest-possible latency over other goals, especially throughput and resource consumption, and (2) which options to look for to implement those changes.

The guide is broken up into three sections:

1. Tuning Your Host Resources, about your hardware and operating system,
2. Tuning Your Ultra Messaging Applications, about your application code and 29West configuration options you should consider in a low-latency context, and
3. Tuning Your Ultra Messaging Sample Applications,  if you wish to use those samples to establish performance benchmarks for your network.

As an example of the type of information we provide, one change that we have found essential to minimizing latency is to disable the “C-state” energy-saving settings found on most servers. These settings do save power, but at the cost of higher latency, due to the time needed to restore full power to processing resources on demand.

Customers can use their 29West KnowledgeBase login to access the 29West Ultra Messaging Low Latency Tuning Guide. If you don’t have a KB login, please contact 29West Support to request one.