Ready for all those slow down questions?


benchtestresultsOne of my hottest topics when talking to people is performance.  Everyone wants their systems to run faster and faster, or at least as fast as possible.  So, when a user’s machine suddenly begins to run slower, the support lines usually light up and we need to know the reasons why.

Microsoft have been one of the first organisations to go public on a slowdown that will be introduced as part of updates that they are currently rolling out to just about every Windows machine.  This is a necessary update to address a very serious CPU vulnerability that could leave sensitive data open to access by hackers.  If you and your users are running the latest Windows 10 updates, then you will most likely not notice much of a change.  It’s said that the percentage slowdown on the latest Windows 10 operating system is in the single percentage digits, so somewhere between 1 and 9%.  When we are talking milliseconds, most users will not notice.

However, for people running older Windows 10 versions, Windows 8, Windows 7 and anyone mad enough to still be running older versions of Windows, the performance hit will almost definitely be noticeable, with talk of performance hits of up to 30%.

Now, that IS some significant hit on our valuable time.

RevUS will shortly be pushing out a formal announcement and I’ll share that with my readers.  In the meantime, I feel that it will be very worthwhile for all technical support and account managers who are involved with supporting and managing OI based systems, to read the following articles:

https://newsstand.google.com/articles/CAIiECBFhLaOKoHY4QkVsN5dgYIqFwgEKg4IACoGCAow3O8nMMqOBjCkztQD

https://www.theguardian.com/technology/2018/jan/04/meltdown-spectre-computer-processor-intel-security-flaws-explainer

I’d like to thank Mike Ruane for giving me a heads up on this important issue that is sure to have our users emailing and phoning us when their OI based systems begin to run slower.  Naturally, the best advice (as always) is to make sure that your systems are being deployed to properly patched Windows operating systems and that you are running OpenInsight 9.4 or later with the Universal Driver 4.7.2 or later.

Advertisements

Things just got a little quicker.


benchtestresultsThis will be my last posting of the day with regards to the new Universal Driver (UD) releases, I promise, no more interruptions to your day from me 😉

Everyone keeps talking to me about performance these days and I wanted to share with you that I have heard a whisper that a test using the UD 4.7.2 (and I believe the UD 5.0.0.4 because that has the same enhancements) is performing better than previous versions during a mammoth testing project.

Testing is on going and loads are increasing, but I’m hearing about 17,000 reads being completed in around 3 seconds.  Obviously different systems will perform differently, but if performance is important to you, then maybe now is the time to take a look at what OpenInsight 9.4 and the Universal Driver 4.7.2 or the 5.0.0.4 can do for you.

UD5 vs UD4.7


benchtestresultsI don’t look at the Knowledge Base on www.revelation.com enough and I’m often surprised at the information that resides there.  Today was no exception and just proves how long it was since I took a look and how important it is to check there fore time to time.

I have been talking to a few people about the UD5 and the upgrade process from earlier versions of the Universal Driver and the benefits of the upgrade and performance ‘always’ comes up.  I was therefore pleasantly surprised to see that Revelation have recently published some bench test results which compare the Universal Driver 4.7 to the Universal Driver 5.0, running both with and without encryption.

There are four tests which were completed using all three configurations (UD4.7 without encryption, UD 5.0 without encryption and UD 5.0 with encryption).  The results are nicely summarised in a graphic and there are also detailed screen shots of the results screens and the parameters used for each of the tests – totally transparent reporting.

I’ll leave you to review the knowledge base article, but it was good to see that the UD 5.0 is consistently faster than the UD4.7 when running without encryption. 

Running with encryption sees a decrease in performance of around 2 to 2.5 times the time taken in the UD 4.7 without encryption.Which is expected for obvious reasons.

Please see www.revsoft.co.uk for more details about the Universal Driver or contact your local Revelation representative or application author to discuss your upgrade to the UD5.

Improving throughput in OpenInsight


As I am non-technical, it is always nice to receive some technical information from around the Revelation community. Revelation often shares things with me, but there is a wealth of technical knowledge and information out there in the community. I watch many of the blogs but a heads up is always useful and Martin Drenovac from Powerforce recently contacted me with a very useful tip:

——————————————————————————————————————

Martyn good evening

Just reading your blogs and OI pointing to the use of cacheing – we’ve recently started to refactor some portions of our system to use rtp65 as a means of improving throughput.

And internally we use the web to document our own code / specs – and I wonder whether this article (not yet complete), might be a useful example to put into the OI community.

http://www.rosterme.com.au/2013/support/faq7.html

Please note that it’s not fully written up just yet.

Cheers,
Martin Drenovac
Powerforce Software P/L

Increasing Application Performance


Did you know…

Application performance is key to modern systems and everyone needs to achieve more, in a faster and faster timeframe and with greater efficiency.  Alot can be achieved through a system healthcheck (hardware and code reviews), the UD4.7 can be fully utilised and other things.

However,  storing frequently accessed values in memory, rather than on disk is a technique for improving your applications performance.  One useful tool is an in-memory hashtable, also known as key/value storage.  The hashtable lets you write a value to a key and read it later.  Several OpenInsight tools offer to help you hold values in memory, but each tool has strengths and weaknesses – so how do you know which tool to use in which situation.

To find out more, Click Here to review Revelation’s Technical Bulletin on OpenInsight In Memory Hashtables.