July 25th, 2016 • Comments Off on Insight driven resource management & scheduling
Future data center resource and workload managers – and their [distributed]schedulers – will require a new key integrate capability: analytics. Reason for this is the the pure scale and the complexity of the disaggregation of resources and workloads which requires getting deeper insights to make better actuation decisions.
For data center management two major factors play a role: the workload (processes, tasks, containers, VMs, …) and the resources (CPUs, MEM, disks, power supplies, fans, …) under control. These form the service and resource landscape and are specific to the context of the individual data center. Different providers use different (heterogeneous) hardware (revisions) resource and have different customer groups running different workloads. The landscape overall describes how the entities in the data center are spatially connected. Telemetry systems allow for observing how they behave over time.
The following diagram can be seen as a metaphor on how the two interact: the workload create a impact on the landscape. The box represent a simple workload having an impact on the resource landscape. The landscape would be composed of all kind of different entities in the data center: from the air conditioning facility all the way to the CPU. Obviously the model taken here is very simple and in real-life a service would span multiple service components (such as load-balancers, DBs, frontends, backends, …). Different kinds of workloads impact the resource landscape in different ways.
(Click to enlarge)
Current data center management systems are too focused on understanding resources behavior only and while external analytics capabilities exists, it becomes crucial that these capabilities need to move to the core of it and allow for observing and deriving insights for both the workload and resource behavior:
- workload behavior: Methodologies such as described in this paper, become key to understand the (heterogeneous) workload behavior and it’s service life-cycle over space and time. This basically means learning the shape of the workload – think of it as the form and size of the tile in the game Tetris.
- resource behavior: It needs to be understood how a) heterogeneous workloads impact the resources (especially in the case of over-subscription) and b) how features of the resource can impact the workload performance. Think of the resources available as the playing field of the game Tetris. Concept as described in this paper help understand how features like SR-IOV impact workload performance, or how to better dimension the service component’s deployment.
Deriving insights on how workloads behave during the life-cycle, and how resources react to that impact, as well as how they can enhance the service delivery is ultimately key to finding the best match between service components over space and time. Better matching (aka actually playing Tetris – and smartly placing the title on the playing field) allows for optimized TCO given a certain business objective. Hence it is key that the analytical capabilities for getting insights on workload and resource behavior move to the very core of the workload and resource management systems in future to make better insightful decisions. This btw is key on all levels of the system hierarchy: on single resource, hosts, resource group and cluster level.
Note: parts of this were discussed during the 9th workshop on cloud control.
March 21st, 2016 • 1 Comment
Orchestration and Scheduling are not the newest topics, in fact they have been used in distributed systems forever (as in a couple of decades :-)). Systems like Mesos and Kubernetes (or offerings like Mantl) have brought advancements when it comes to dealing with scale. Other systems have a great background in scheduling and offer many (read a whole lot) policies for the same, this includes technologies like Grid Engine, LSF/OpenLava, etc.. Actually some of these technologies integrate with each other (like navops, Kubernetes and Mesos, OpenLava and Mesos, …), which makes it for example interesting when dealing with scheduling for space & order at the same time.
Next to pure demand, upcoming trends like CNCF & OCI as well as the introduction of Software Defined Infrastructure (SDI) drive the number of resources and services the Orchestrators and Controllers manage up. And the Question arises how to efficiently manage your data center – doing it by a human pressing a button is just not going to scale 🙂
Feedback control systems are a great start, however have some drawbacks. The larger the scale the more conflicts you might get between the feedback loops. The approaches might work up to rack level but probably not much beyond that. For large scale we need an approach which works along the lines of watch (e.g. by using snap), learn/decide (e.g. by using TAP) and act (See Jason Waxman’s keynote at OCP). This will eventually allow for a operatorless/humanless/driverless operations of the data center to support autonomous operations for scaling, healing and optimizing e.g. TCO.
Within Intel Labs we have therefore come up with the concept of a foreground and a background flow. Within a continuously running background flow we observe (if needed over long time-periods) the data center with its resources and services and try to derive & update models heuristics (read: rule of thumb) continuously using analytics/machine learning. Within a foreground flow – which sometimes is denoted the fast loop as it needs to perform – we can than score against those heuristics/models in actions plans/recipes.
The action plan/recipes describe a process on how we deal with a initial placement or re-balancing event. The scoring will allow for making better initial placement (adding a workload) as well as re-balancing decisions (how/what/when to kill, migrate or tune the infrastructure). How to derive an heuristics is explained in a paper referenced below – the example within that is about to learn how to best place a VNF so that is makes optimal use of platform features such as SR-IOV. Multiple other heuristics can easily be imagined, like learning how many cores a certain workload needs.
The following diagram shows the background and foreground flow.
(Click to enlarge)
The heuristics are stored in an Information Core which based on the environment it is deployed in tunes itself. We’ve defined the concepts described here in a paper submitted to the Middleware 2015 conference. The researchers from Umea (who also run this highly recommended workshop) have used it and demonstrate an example use case in the same paper. For an example on how a background flow can help informing the foreground flow read this short paper. (Excuses for the paywall :-))
I’ll follow-up with some more blog posts detailing certain aspects of our latest work/research, like how the landscape works.
October 30th, 2014 • Comments Off on American Football Game Analysis
I’ve been coaching American Football for a while now and it is a blast standing on the sideline during game day. The not so “funny” part of coaching however – especially as Defense Coordinator – is the endless hours spend on making up stats of the offensive strategy of the opponent. Time to save some time and let the computer do the work.
I’ve posted about how you could use suricate in a sports data setup past. The following screen shot show the first baby steps (On purpose not the latest and greatest – sry 🙂 ) of analyzing game data using suricate with python pandas and scikit-learn for some clustering. The 3D plot shows Down & Distance vs Run/Pass plays. This is just raw data coming from e.g. here.
The colors of the dots actually have a meaning in such that they represent a clustering of many past plays. The clustering is done not only on Down & Distance but also on factors like field position etc. So a cluster can be seen as a group of plays with similar characteristics for now. These clusters can later be used to identify a upcoming play which is in a similar cluster.
(Click to enlarge)
The output of this python script stores processed data back to the object store of suricate.
One of the new features of suricate is template-able dashboards (not shown in past screenshot). Which basically means you can create custom dashboards with fancy graphics (choose you poison: D3, matplotlib, etc):
(Click to enlarge)
Again some data is left out for simplicity & secrecy 🙂
Making use of the stats
One part is understanding the stats as created in the first part. Secondly acting upon it is more important. With Tablets taking on sidelines, it is time to do the same & take the stats with you on game day. I have a simple web app sitting around in which current ball position is entered and some basic stats are shown.
This little web application does two things:
- Send a AMQP msg with the last play information to a RabbitMQ broker. Based on this new message new stats are calculated and stored back to the game data. This works thanks to suricate’s streaming support.
- Trigger suricate to re-calculate the changes of Run-vs-Pass in an upcoming play.
The webapp is a simple WSGI python application – still the hard work is carried out by suricate. Nevertheless the screenshot below shows the basic concept:
December 23rd, 2013 • Comments Off on Live football game analysis
(Click to enlarge)
Note: I’m talking about American Football here 🙂
In previous posts I already showed how game statistics can be used to automatically which Wide receiver is the Qb’s favorite on which play, down and field position. Now let’s take this one step further and create a little system (using Suricate) which will make suggestions to a Defense Coordinator.
The following diagram will guide through the steps needed to create such a system:
Live game breakdown (Click to enlarge)
Let’s start at the top. (Step 1) The User of Suricate will start with performing some simple steps. First a bunch of game statistics are uploaded (same as using in this post). Next also a stream is defined. In this case a URI for an AMQP broker (using CloudAMQP – RabbitMQ as a Service) is defined in the service. With this used defined data is provide to the service.
(Step 2) Now we start creating an analytics notebook. Suricate provides an interactive python console via your web browser which can easily be used to explore the data previously uploaded. Python Pandas and scikit-learn are both available within the Suricate service and can be used right away to accomplish this task:
Exploring game statistics (Click to enlarge)
Based on the data we can create a model which describes on which down, on which fieldposition a run or a pass play is performed. We can also store who is the favorite Wide receiver/Running back for those plays (see also). All this information is stored in a JSON data structure and saved using the SDK of Suricate (Step 3).
(Step 4) Now a little external python script needs to be written which grabs relatively ‘live’ game data from e.g. here. This script now simple continuously sends messages to the previously defined RabbitMQ broker. The messages contain the current play, fieldposition and distance togo information.
(Step 5) Now a processing python notebook needs to be written. This is a rather simple python script. It takes the new incoming messages and compares them to the model learned in step 2 & 3. Based on that suggestions can be displayed (Step 6a) – “e.g. watch out for Wes Welker on 3rd and long at own 20y line” or just some percentages for pass or run plays:
Processing notebook (Click to enlarge)
Next the information about the new play can be added to the game statics data file (Step 6b). Once this is done a new model can be created (Step 7) to get the most up to date models all the time.
With this overall system new incoming data is streamed in (continuous analytics), models updated and suggestions for a Defense Coordinator outputted. Disclaimer: some steps describe here are not yet in the github repository of suricate – most namely the continuously running of scripts.
November 17th, 2013 • 1 Comment
Again a little excerpt on stuff you can do with the help of Suricate. Again we’ll look at play-by-play statistics. So no information on which play was performed, but just the outcomes of the plays. Still there is some information you can retrieve from that. Most importantly because it can be done automated without user interaction needed. Just upload the file, press a button and get the results:
Peyton’s favorites (Click to enlarge)
This time you are looking are a cluster analysis of players Peyton passed too in the game of week 1. First cluster represent players passed to on 1,2 & 3 down with up to 6 yards. The second cluster the goto-guys which did go for medium yardage and finally the WOs able to get a bunch of yards on the board.
So with simple scripts (few lines of code – which can be reused) it is possible to abstract information from just play-to-play statistics. I guess mostly important to Defense Coordinators who would love to get some information on the fly with the press of a button 🙂