Tags / Netkernel
External Request ProfilerWe’ve been working recently with a client, on a solution, revolving around crunching large volumes of data. The scale of data meant that even on a large cluster the computation takes many hours. The usual mode of operation for these compute jobs is to set them off in the evening, and return the next working day to see the results. I was working with them to capture some engineering figures and statistics to assess the performance and opportunities to optimise or re-design.
Published Jan 26, 2018 3 min read netkernel
Generalizing Dynamic Programming
- Using dynamic programming techniques to optimise general compute efficiencyDynamic programming is an approach to efficiently implementing certain classes of algorithms. As a feeble excuse for not noticing this earlier, the term is a little confusing: it isn’t about programming or about doing it dynamically! It was invented in the 1950s by Richard E. Bellman whilst working for RAND Corporation. Apparently he deliberately named it obtusely to avoid his work being considered research! I discovered this term recently when looking into existing research about algorithm optimisation and specifically why the systemic memoisation in NetKernel can optimise solutions to certain problems particularly well.
Travelling Salesman Problem
- Study of TSP running on NetKernel with cache replacing dynamic programmingThis paper offers an analysis of the application of NetKernel and Resource oriented computing to the well known Travelling Salesman Problem (TSP). This discussion ignores the many approximate approaches to solving this problem which are necessary for real-world problems of any significant complexity. The aim, however, is to show how ROC can automatically optimise a solution which it knows nothing about other than empirical data collected during execution. The algorithm structures the problem into sub-problems many of which overlap with each other - this occurrence is called overlapping subproblems.
Practical State Machines in NetKernel
Following up on my previous post, ROC Hockey, where I introduced our new implementation of Hierarchical State Machines on NetKernel I’ve decided to cover the topic of how they are actually used in practice in an ROC system.
NetKernel IDE FeedbackThank you to all those who provided feedback on what you’d like to see in a NetKernel IDE Plugin. This is a topic that has been on my mind for a couple of years now as various folks have dabbled with creating one, and as we have observed many development teams and how they use their IDEs to truly integrate all their tools. So the feedback I received was pretty much in consensus in some areas and widely varied in others.
Published Apr 12, 2013 3 min read netkernel
Heap Tuning for Large NetKernel Instances Part 2In the last post I set the stage for a discussion on how to tune large NetKernel instances by providing a discussion of how the Java heap operates and how the NetKernel representation cache interacts with that. Now we are ready to get into the details of how to tune your system. So let’s get stuck in. First we need to capture a of couple metrics: Fixed OldGen - this is composed of the long lived objects that are the modules, spaces, endpoints and kernel data structures.
Heap Tuning for Large NetKernel Instances Part 1I’m writing this post to document information that I learned whilst developing the new NetKernel enterprise L1 representation cache (released today!). Knowledge of how to tune large Java instances is readily available but because NetKernel uses a cache as an integral part of it’s operation and usually this takes up a significant majority of the heap space of a running system this effects the conventional wisdom. So this article summarizes and augments this information with tips and details for tuning large NetKernel instances.
Cut to the Chase“Cut to the chase” is something you’re only likely to hear me say if I’m enduring some kind of Dickensian wordfest drawn out from the changing of a single bit of information. However in this case I’m not the actor but rather it’s a new architectural endpoint for NetKernel that can keep your web apps on track and up-to-date. Let me explain. Whilst working on the recent representation cache update I noticed that the standard pattern of issuing AJAX requests from the browser was causing a problem.
New Enterprise Representation CacheLast week we had a get together with representatives of some of our key partners in Brussels. There, amongst other things, I demonstrated a key new technology that I’ve been working on for the last couple of months. This is an all new enterprise grade representation cache. The NetKernel representation cache is a mechanism for storing the responses from endpoints with the aim of eliminating duplicate computation. This is made possible because every computation is uniquely identified with a combination of resource identifier and request scope.
Dynamic GrammarsLast week I received a request for advice on an interesting problem from Grégoire Colbert. He’s given me permission to share his request and I hope to share with you the solution. Let’s imagine that I have a constantly evolving OWL file of living species. I want NetKernel to read this OWL resource, and allow the user to get a representation for each entry in the OWL file : > /birds/sparrow > /fishes/tuna The difficulty is that when NetKernel starts, I (as a programmer) don’t have any idea of what the OWL file will contain, and so I cannot write a static grammar with /birds/sparrow and /fishes/tuna.
Published Sep 28, 2012 5 min read netkernel