Cloud Computing Benchmark
RB-A, the 1st
step to continuous price-performance benchmarking
of the cloud
Edward Wustenhoff, ...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 2
Table of Contents
INTRODUCTION...........................
Cloud Computing Benchmark 3
Introduction
Consumer Internet businesses, like eBay, Twitter and Facebook depend on their com...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 4
changes in a 24-month period. And as we have observed ...
Cloud Computing Benchmark 5
Methodology
When we started thinking about what was different between the new performance
dyna...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 6
Process
Figure 1: RB-A Test Process shows the high-lev...
Cloud Computing Benchmark 7
Some required us to open up a separate account for other countries, all interesting
indicators...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 8
Scope
The below table shows the scope of this initial ...
Cloud Computing Benchmark 9
We are pursuing an equivalent test for Windows and possibly other operating systems.
The resul...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 10
Benchmark Results
Summary
Below are some highlights f...
Cloud Computing Benchmark 11
 Not all locations are created equal in availability and performance of instance types.
The ...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 12
Definitions
In order to better understand the graphs ...
Cloud Computing Benchmark 13
Performance by Cloud Service Provider
Performance of 1 core instances can vary by 622% betwee...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 14
Figure 3: Google Compute Engine
Figure 4: HP Cloud
Cloud Computing Benchmark 15
Figure 5: Linode
Figure 6: Microsoft Azure
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 16
Figure 7: Rackspace
Figure 8: Softlayer
Amazon AWS ha...
Cloud Computing Benchmark 17
We note a lot of cloud service providers have similar performance scores for different
instan...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 18
You can see an example of how that looks like for 1 c...
Cloud Computing Benchmark 19
Absolute Performance
The top three highest performing (average over 15 days) cloud services o...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 20
Price performance
Price performance for a 4-core comp...
Cloud Computing Benchmark 21
The top three cloud services for 4 core instances by price-performance (averaged over 15
days...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 22
The best price-performance 4-core compute cloud servi...
Cloud Computing Benchmark 23
Even within a provider the economic impact can be a key differentiator:
Figure 13: Rackspace ...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 24
Performance over time
The same instance performance c...
Cloud Computing Benchmark 25
Figure 15: Google Performance over time
Figure 16: HP Cloud Performance over time
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 26
Figure 17: Linode Performance over time
Figure 18: Mi...
Cloud Computing Benchmark 27
Figure 19: Rackspace Performance over time
Figure 20: Softlayer Performance over time
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 28
As you can see performance is generally stable over t...
Cloud Computing Benchmark 29
To show another perspective we looked at 2 vendors compared:
The below shows Amazon AWS and G...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 30
Global observations
Not all locations are created equ...
Cloud Computing Benchmark 31
Figure 24: Google regional performance
Figure 25: HP Performance by region
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 32
Note: HP Cloud has no presence outside of the US.
Fig...
Cloud Computing Benchmark 33
Figure 28: Rackspace performance by region
Figure 29: Softlayer performance by region
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 34
Although the differences between regions are not extr...
Cloud Computing Benchmark 35
The data for both Google and Azure shows that there is a 100% “failure to launch” rate at
the...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 36
Conclusions
This first comprehensive and continuous p...
Cloud Computing Benchmark 37
Appendix 1: Test details
Test details
Burstorm used the standard UnixBench score but scaled i...
Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 38
Appendix 2: What’s next
We normalized the values agai...
Cloud Computing Benchmark 39
Review History
We’d like to thank all the reviewers below as well as those who chose to remai...
of 39

IaaS Price performance-benchmark

This benchmark is the result of the collaboration between Burstorm and Rice University and uses a high degree of automation. The scope of the first benchmark is seven suppliers across three continents with a total of 96 different instance types. The benchmark was executed every day, for at least 15 days. The results are normalized to a monthly pricing model to establish the price-performance metrics.
Published on: Mar 4, 2016
Published in: Technology      
Source: www.slideshare.net


Transcripts - IaaS Price performance-benchmark

  • 1. Cloud Computing Benchmark RB-A, the 1st step to continuous price-performance benchmarking of the cloud Edward Wustenhoff, CTO, Burstorm T. S. Eugene Ng, Associate Professor, Rice University ABSTRACT This benchmark is the result of the collaboration between Burstorm and Rice University and uses a high degree of automation. The scope of the first benchmark is seven suppliers across three continents with a total of 96 different instance types. The benchmark was executed every day, for at least 15 days. The results are normalized to a monthly pricing model to establish the price-performance metrics.
  • 2. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 2 Table of Contents INTRODUCTION......................................................................................................................................................3 METHODOLOGY......................................................................................................................................................5 PROCESS.................................................................................................................................................................................... 6 SCOPE........................................................................................................................................................................................ 8 BENCHMARK RESULTS ......................................................................................................................................10 SUMMARY ...............................................................................................................................................................................10 DEFINITIONS ..........................................................................................................................................................................12 Performance ...........................................................................................................................................................................12 CPU performance..................................................................................................................................................................12 IO performance......................................................................................................................................................................12 Price............................................................................................................................................................................................12 Price-Performance...............................................................................................................................................................12 PERFORMANCE BY CLOUD SERVICE PROVIDER................................................................................................................13 Absolute Performance........................................................................................................................................................19 PRICE PERFORMANCE...........................................................................................................................................................20 PERFORMANCE OVER TIME..................................................................................................................................................24 GLOBAL OBSERVATIONS.......................................................................................................................................................30 CONCLUSIONS .......................................................................................................................................................36 APPENDIX 1: TEST DETAILS ............................................................................................................................37 TEST DETAILS ........................................................................................................................................................................37 BCU: SYSTEM SPECS ............................................................................................................................................................37 APPENDIX 2: WHAT’S NEXT.............................................................................................................................38 REVIEW HISTORY................................................................................................................................................39
  • 3. Cloud Computing Benchmark 3 Introduction Consumer Internet businesses, like eBay, Twitter and Facebook depend on their computing infrastructure (compute, storage, data centers and networks) as the foundation of their enterprise. Increasingly this is true across other industries including high tech, financial services, biotech, healthcare, etc. These infrastructure components are more and more consumed as a service (Cloud computing). Given the increasing complexity of cloud deployments, Burstorm in 2015 launched the industry’s first Computer-Aided Design (CAD) application for cloud architects. Like Autodesk in construction, Burstorm’s application allows architects to develop new infrastructure designs as well as remodel existing compute, storage datacenter and network infrastructures. The cornerstone of the application is a product catalog, which as of the writing contains over 900 product sets totaling over 36000 products. The product catalog today contains product specifications, pricing covering different types of business models and location information. Based on this product catalog and a class of optimization algorithms the application aids the architect in making design decisions. Over the past several years, Dr. T. S. Eugene Ng’s group at Rice University has also been focused on cloud computing. One of the areas of research interest has been the performance of compute and storage cloud services. Recently they published their joint work with Purdue University: Application-Specific Configuration Selection in the Cloud: Impact of Provider Policy and Potential of Systematic Testing1, in the Proceedings of IEEE INFOCOM'15. The paper takes a first step towards understanding the impact of cloud service provider policy and tackling the complexity of selecting configurations that can best meet the price and performance requirements of applications. Their work sparked the interest of Edward Wustenhoff at Burstorm. At the same time, Dr. Ng was hoping to collaborate with practitioners to get exposed to a wider set of configuration choices, and other compute & storage cloud service providers, beyond Amazon EC2. There are a number of challenges to price-performance benchmarking since Jim Gray’s landmark paper, A Measure of Transaction Processing Power2. First, as Burstorm’s product catalog shows there are now 1000s of different compute & storage cloud services. These cloud services span many different locations. One might think it’s odd to talk about location and cloud services in the same sentence, but for geopolitical and networking performance reasons the location of these cloud services does matter. Furthermore, there are a variety of business models. One can see several cloud price 1 Mohammad Hajjat, Ruiqi Liu, Yiyang Chang, T. S. Eugene Ng, Sanjay Rao, "Application-Specific Configuration Selection in the Cloud: Impact of Provider Policy and Potential of Systematic Testing” in Proceedings of IEEE INFOCOM'15, Hong Kong, China, April 2015. 2 Jim Gray, A Measure of Transaction Processing Power, 1985 http://www.hpl.hp.com/techreports/tandem/TR-85.2.pdf
  • 4. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 4 changes in a 24-month period. And as we have observed the performance of the same instance can be different at different times and different locations. On top of that you can consume services by the hour, month, annually or one can buy on the spot market. New products are being introduced on a monthly basis and pricing can change weekly. For instance, Amazon made tens of price changes in a 24-month period. And as the INFOCOM'15 study has observed, the performance of the same instance can be different at different times and different locations. The result of the collaboration with Rice is the industry’s first comprehensive and continuous price-performance benchmark. Using a high degree of automation the scope of the first benchmark is seven suppliers (Amazon, Google, Microsoft, Rackspace, IBM, HP and Linode), across three continents (Asia, North America and Europe) with a total of 266 compute products spread over 3 locations per vendor, where available. The benchmark was executed every day, for 15 days. The results are normalized to a 720-hour, monthly pricing model to establish the price-performance metrics. Most of us are familiar with traditional performance testing. However we believe that those practices are only partially applicable to understanding cloud computing performance. What makes this report unique and interesting is that we tested a large amount of instance types (96) over time, in multiple locations and include economic impact data. Some of the results show a 622% variation of performance within a same instance type. The best performing instance does not show the best price-performance. Availability and behavior of instances is not the same depending on location within the same provider. All in allthe cloud is a very dynamic and complex environment. This PDF report shows selected screen shots from the interactive report, which is available by subscribing at http://www.burstorm.com/rba. The interactive report allows you to visualize the data in many different ways, and contains results of the ongoing continuous benchmarking. Our plans for the future include: adding more cloud service providers, increasing the number of locations, providing benchmarking services for private cloud services and development of RB-B. More forward looking statements can be read in Appendix 2: What’s next. Finally, while this data is important and can be used to refine and improve product offerings, Burstorm has also plans in progress to incorporate the data into its CAD application so cloud architects can create architectures optimized for locality, price, performance and price-performance.
  • 5. Cloud Computing Benchmark 5 Methodology When we started thinking about what was different between the new performance dynamics in cloud computing and the TPC Benchmark days, we realized that because there is less control over the environment we cannot assume that every instance tested is identical at start up, over time and per location. This causes great uncertainty about the capability to process workloads consistently. In addition all the new business models raise questions about the economic benefits for certain instance types. Selecting the optimal instance for a specific workload has become also a function of performance and economics. We came to the realization that the only conclusive way to address this, would be by continually testing all instance types everywhere. One can imagine how this becomes a logistical and economic challenge that is seemingly impossible to address. It’s because of this we came to change how we think of the benchmarking process and not so much about creating a better benchmark. Of course certain aspects within a Virtual Machine will need to be tested differently and Burstorm is working with Rice University on improving the benchmarks, but in the end the biggest challenge is around scale and velocity. Fortunately in the new compute era, the time and cost to create a benchmark environment can be measured in cents and minutes and is easily distributed through automation. The process and scope we applied are outlined below.
  • 6. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 6 Process Figure 1: RB-A Test Process shows the high-level process we use to spin up, benchmark, write results and display the results. Figure 1: RB-A Test Process The basic concept is to spin up instances, run the benchmark, write the data to the Burstorm product catalog, combine it with our pricing data and repeat this for about 15 days for each instance type, for each provider at each selected locations. Because not all providers in the target set have services in the same locations we decided to select one for each provider in Asia, the US and Europe so we could spot potential differences in deployment and cost per region. We intend to expand providers and locations as we continue our benchmarking. It was interesting to experience the difference in deployment processes for each supplier. Some contacted us to ensure we were legitimate, others wanted financial guarantees and one would send us several emails for each instance type started to confirm approval, actual provisioning and a “Getting started” email. We also found some bugs in provider’s deployment API’s we had to fix before we could proceed.
  • 7. Cloud Computing Benchmark 7 Some required us to open up a separate account for other countries, all interesting indicators of the maturity level of this market place. Benchmarks were run in parallel, though with some damping to avoid limits (cpu, memory etc.) of various providers. The instances were created using standard chef knife CLI commands (e.g. "knife [provider] server create"), which started and loaded the benchmark software onto the instance. When finished, the software reported back the test results to our server using a JSON version of the standard UnixBench test results. Due to the scale and need for automation, we used best effort to gather the data for each daily run, and as such there are sometimes missing data points. We allowed for this as opposed to trying to fix failures to launch because it is another interesting data point. However, since this is out of scope for this report, we haven't diagnosed deeply why the instances we tried to spin up didn't start, but some portion of it seems to hint towards capacity limitations on the provider's side because missing instance runs were often in the larger 8-16 core variety. Burstorm uses the standard UnixBench score but scaled to a more modern processor and bus (a Raspberry Pi 2, ARM7 @900Mhz) from the original SparcStation 20-61 "George". The detailed specifications of the system and tests can be found in Appendix 1: Test details The last phase was to create the views of the data, the main content of this report. We combined the performance data with the product pricing catalog data from Burstorm’s CAD application to create the price-performance benchmark numbers. The fact that this performance data is now in context of our CAD application, which amongst other things includes design capabilities and private product sets, opens up the opportunity to expand into benchmarking private cloud services and multi system designs in the future.
  • 8. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 8 Scope The below table shows the scope of this initial project creating 7839 test results each with multiple data points. See Appendix 1: Test details for how the data points were created. Note that we confined the number of locations to three for the reasons mentioned earlier and note that the total number of test results is less than the exact math due to not running 1 and multi core tests for 1 core systems and some “failures to launch”. Due to a tooling issue we want to recognize that we did not include some key instance types from Rackspace. We have since corrected the issue and we expect them to have noticeably different result in the interactive report. For this report we did not test any dedicated or “bare metal” type instances. These will be added in the future. Provider # Instance Types # Locations # Products AWS 30 3 90 Google 14 3 42 Rackspace 9 3 27 Azure 18 3 54 Linode 9 3 27 HP 11 1 11 Softlayer 5 3 15 Selected total 96 19 266 Table 1: Testing Scope The following locations were selected for each region by provider: Provider North America (NA) Europe (EMEA) Asia (APAC) AWS Ashburn US Dublin IE Singapore SG Google Council Bluffs US Saint-Ghislain BE Changhua County TW Rackspace Grapevine US Slough GB Hong Kong HK Azure California, CA Omeath IE Singapore SG Linode Fremont US London GB Singapore SG HP Tulsa US N/A N/A Softlayer San Jose US Amsterdam NL Singapore SG Table 2: Locations by provider We did not separately test Windows instances for the following reasons:  Not all providers have windows instances and we wanted to make sure we had a common baseline.  Assuming the impact of the underlying virtualization to be equal for any OS, we expect the relative performance between 4core and 8core systems to be somewhat equal for both Windows and Linux.
  • 9. Cloud Computing Benchmark 9 We are pursuing an equivalent test for Windows and possibly other operating systems. The result is the industry’s first continuous and comprehensive price-performance benchmark. The following results are just one view of the data, which can be seen in many different views from the interactive report
  • 10. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 10 Benchmark Results Summary Below are some highlights from the results of our approach:  Performance of 1-core instances can vary by 622% between providers.  Price performance for a 4-core compute cloud service can vary by 1000%.  The same instance performance can fluctuate by 60% over time
  • 11. Cloud Computing Benchmark 11  Not all locations are created equal in availability and performance of instance types. The rate of change in instance types, pricing, performance over time and availability of services by location confirms that the traditional way of benchmarking a small set of instance types in a unique event is not sufficient anymore in today’s world of cloud computing. Even within the short span of this report, Google updated their infrastructure and pricing. To keep up with this is precisely why we created the interactive report. The next chapters show the details and data that led up to these findings. But let’s provide you with some definition of terms first.
  • 12. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 12 Definitions In order to better understand the graphs and statements made, below are the definitions of the key metrics used in this and the interactive report. Performance This reflects the UnixBench score relative to the Burstorm Compute Unit (BCU) baseline. See more on that in Appendix 1: Test details. When multiple data points applied an average of the scores was taken. A higher score means better performance. CPU performance CPU performance is measured using a subset of the UnixBench tests, namely: 1. dhry2reg -- Dhrystone CPU using two register variales 2. whetstone-double -- Whetstone double precision CPU test 3. pipe -- Unix pipe throughput 4. context1 -- Pipe based context switching throughput 5. shell8 -- 8 bash shells executing simultaneously A higher score means better CPU performance. IO performance IO performance is measured using a subset of the UnixBench tests, namely: 1. fstime -- file copy, 1024 byte buffer size, 500 maxblocks 2. fsbuffer -- file copy, 256 byte buffer size, 500 maxblocks 3. fsdisk -- file copy, 4096 by buffer size, 8000 maxblocks A higher score means better IO performance. Price Price is the monthly cost using hour-hour terms, normalized to 720/hrs/month, no prepayments and using Ubuntu 14.04 Linux. The prices used in this document reflect the prices of the instance running the specified OS at the start of the test period. Realize that a Redhat or Windows OS instance types would typically carry a higher price. Price-Performance Price performance is defined as the instance score divided by the monthly cost for the instance. A higher score means better price-performance.
  • 13. Cloud Computing Benchmark 13 Performance by Cloud Service Provider Performance of 1 core instances can vary by 622% between providers. The first view of the benchmark results look at the range of performance by cloud service provider. The details of how the numbers were generated can be found in Appendix 1. The X-axis is the instance type, the Y axis is the relative performance against the Burstorm Compute Unit (BCU) calculated as an average over all available data points. Figure 2: Amazon AWS
  • 14. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 14 Figure 3: Google Compute Engine Figure 4: HP Cloud
  • 15. Cloud Computing Benchmark 15 Figure 5: Linode Figure 6: Microsoft Azure
  • 16. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 16 Figure 7: Rackspace Figure 8: Softlayer Amazon AWS has the largest variety of options followed by Azure and Google. Amazon AWS also has the highest performing instances. The interactive report will allow you to compare different suppliers, and different instance types.
  • 17. Cloud Computing Benchmark 17 We note a lot of cloud service providers have similar performance scores for different instance types. We believe this to be a function of 2 variables that play here: The UnixBench performance score does not show a lot of impact from different memory sizes and the differences in IO capability of the instance type. The latter becomes clearer when you look at Amazon AWS CPU scores vs IO scores: IO shows a more linear pattern. Figure 9: AWS CPU vs IO performance This is also where we want to point out that Amazon AWS has a T-series instance type that has a “performance quota”. This means that as you use the instance over time you use up the quota and once used up, the performance goes down. This favors our testing method where we run 1 benchmark per instance 1x per day in less than 30 minutes, as opposed to continually testing 1 instance over a longer period of time.
  • 18. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 18 You can see an example of how that looks like for 1 core systems in the picture below: Figure 10: performance scores for 1 core instances The revealing part is that as a result of the diversity of platforms and solutions, our benchmark shows scores between 1.8 and 11.2 or a 622% performance difference between the lowest and highest performing 1 core instances. You can also see the difference between 2,4, 8, 16, 32 and 36-core instances in the interactive report.
  • 19. Cloud Computing Benchmark 19 Absolute Performance The top three highest performing (average over 15 days) cloud services of all are:
  • 20. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 20 Price performance Price performance for a 4-core compute cloud service can vary by 1000% The Burstorm CAD application’s product catalog contains product pricing and so we were able to connect a price to each of the instances. While the Burstorm application’s product catalog contains many pricing models (hourly, month-to-month, 12, 24, 36 months etc.) in these results we used the hourly rate without discounts. The interactive report will contain other pricing models in the future. Figure 11 shows the performance and price-performance of all 4-core compute cloud services from the seven suppliers. Price-performance scores are calculated by dividing the performance score by the price/month. The lowest score is .023 and the highest is .23 representing a 1000% difference. Figure 11: Price Performance of 4 core instances
  • 21. Cloud Computing Benchmark 21 The top three cloud services for 4 core instances by price-performance (averaged over 15 days) are:
  • 22. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 22 The best price-performance 4-core compute cloud services is the Linode-4GB (0.23) and is ten times better than the price-performance cloud service from Rackspace’s performance2- 15 (0.023). In fact, the below graph shows that the Linode-4GB is even nearly two times better than the number two, Amazon’s c4.xlarge. While this paper is being finalized Linode is upgrading their virtualization layer and we included more Rackspace instances. The effects are already visible in the continuous benchmark and are available in the interactive report. Figure 12: Price performance for 4-core instance types As you can see, normalizing to price performance can significantly change the picture.
  • 23. Cloud Computing Benchmark 23 Even within a provider the economic impact can be a key differentiator: Figure 13: Rackspace 4 core systems You can see how similar systems from a performance perspective (including IO and CPU) have almost a 500% difference in price performance. The ability to see this impact helps ask questions of what is really different and relevant. If your workload can be distributed over multiple instances looking at price performance is critical to finding the right instance for you. Prices change regularly so this is something you want to monitor over time and adjust accordingly. Since we bind the price to the data point in time the interactive report will soon show not only the current price-performance but also how the price-performance changes for an instance type over a period of time.
  • 24. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 24 Performance over time The same instance performance can fluctuate by 60% over time As we said in the introduction the benchmark was run every day for 15 days and we are continuing to benchmark. As you can see in the detail, performance of a particular compute cloud service can vary over time. In this case there was no change in pricing captured over the fifteen-day period, but we have seen changes on a monthly basis. The next charts show the changes in performance over time by cloud service provider. Each data point in time is the average of all locations performance results for that instance type. Figure 14: AWS performance over time
  • 25. Cloud Computing Benchmark 25 Figure 15: Google Performance over time Figure 16: HP Cloud Performance over time
  • 26. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 26 Figure 17: Linode Performance over time Figure 18: Microsoft Azure Performance over time
  • 27. Cloud Computing Benchmark 27 Figure 19: Rackspace Performance over time Figure 20: Softlayer Performance over time
  • 28. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 28 As you can see performance is generally stable over time but still could vary by as much as 60% within a single instance type (see figure 21). Generally the CPU volatility is less than the IO volatility and volatility looks worse at the higher performing instances. Figure 21: IO over time, 4Core GCE instances A secondary observation is that “failure to launch” events (and thus missing data points) seem to happen more at the larger instance types.
  • 29. Cloud Computing Benchmark 29 To show another perspective we looked at 2 vendors compared: The below shows Amazon AWS and Google Compute Engine 4 core instances over the 15 days tested. Note that the Google instances are represented by the bottom 3 lines. Figure 22: 4 core AWS vs Google over time Amazon AWS looks consistent but more interesting is the approximate 20+% improvement of the Google Compute Engine instances over time. We confirmed that Google upgraded their infrastructure during our test period. Such events happen all the time and confirm we need to gather and look at benchmark data in a different way than a single point in time. For those of you interested, the interactive report has already more data points than is represented in this report. You can see that performance over time matters and it shows that there are significant differences that can impact what the ideal profile for a specific workload is, based on time and day.
  • 30. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 30 Global observations Not all locations are created equal in availability and performance of instance types We spread the testing over 3 locations for each provider in 3 geographies: NA (North America), APAC (Asia) and EMEA (Europe) to benchmark performance and price- performance based on locality. Note that the results are and average of all data points collected. If no data points were collected it means it had a 100% failure to start which most often means not available but could mean a systemic failure in tooling. We are continually working with suppliers to diagnose the root cause. Here are the screenshots from the interactive report: Figure 23: Amazon AWS regional performance
  • 31. Cloud Computing Benchmark 31 Figure 24: Google regional performance Figure 25: HP Performance by region
  • 32. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 32 Note: HP Cloud has no presence outside of the US. Figure 26: Linode performance by region Figure 27: Microsoft Azure performance by region
  • 33. Cloud Computing Benchmark 33 Figure 28: Rackspace performance by region Figure 29: Softlayer performance by region
  • 34. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 34 Although the differences between regions are not extreme they are noticeable. Not all instances are available everywhere. Not all high performance Microsoft Azure instance types are available in APAC for example and HP has only cloud services in the US. When you take a closer look you can find evidence that not all regions are created equal. Comparing 8 Core compute cloud services of Microsoft Azure and Google Compute Engine shows: Figure 30: 8 cores by region: Azure vs Google Microsoft Azure has more variety in performance but the best performing instance is not available in APAC. With the Google Compute Engine cloud services, EMEA has consistent higher performance than North America or Asia Pacific.
  • 35. Cloud Computing Benchmark 35 The data for both Google and Azure shows that there is a 100% “failure to launch” rate at the higher performing instance types in some of the tested locations: Figure 31: Google Compute Engine and Microsoft Azure performance by region At Google the n1-standard-16 is not available in the US and the n1-higmem-16 instance type is not available in EMEA and APAC. At Microsoft Azure the A8 and A9 are not available in APAC. The data shows there are performance differences between regions for the same instance types within a vendor and not all instances are available in every region. Our plans for the interactive report are to increase the number of locations we benchmark.
  • 36. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 36 Conclusions This first comprehensive and continuous price-performance benchmark has yielded some important observations:  Performance of 1-core instances can vary by 622% between providers.  Amazon took the top three absolute performance spots with the c4.8xlarge, c3.8xlarge d2.8.xlarge.  Price performance for a 4-core compute cloud service can vary by 1000%.  The top three price-performance winners for 4 core systems were Linode-4GB, Amazon c4.xlarge and Google n1-highcpu-4.  The same instance performance can fluctuate by 60% over time. The performance over time showed us that a provider can have a bad day and deliver 60% less IO performance capability than on a good day. This highlights how applications will have to learn how to compensate for this volatility if we want to provide more consistent solutions. It also shows the need for continually reporting these numbers to see if the performance gets worse over time due to capacity issues or are getting better due to optimizations and hardware upgrades.  Not all locations are created equal in availability and performance of instance types. Even though most performance and value scores were marginally, but noticeably, the difference in availability of services was significant. HP all instances are only available in the US, At Google the n1-standard-16 is not available in the US and the n1-higmem-16 instance type is not available in EMEA and APAC. At Microsoft Azure the A8 and A9 are not available in APAC. The rate of change in instance types, pricing, performance over time and availability of services by location confirms that the traditional way of benchmarking a small set of instance types in a unique event is not sufficient anymore in today’s world of cloud computing. Even within the short span of this report, Google updated their infrastructure and pricing. To keep up with this is precisely why we created the interactive report. Continuous and comprehensive benchmarking of existing and new cloud services will reveal useful information for both suppliers and consumers of compute & storage cloud services. Rice and Burstorm intend to continue to expand the scope of the benchmarking and work with enterprises, academics and cloud service providers to add to our collective understanding of the cloud.
  • 37. Cloud Computing Benchmark 37 Appendix 1: Test details Test details Burstorm used the standard UnixBench score but scaled it to a more modern processor instead of the original SparcStation. The tests its self were not altered for this version of the benchmark so as to establish a widely understood and vetted baseline. UnixBench is the original BYTE UNIX benchmark suite, updated and revised by many people over the years. The purpose of UnixBench is to provide a basic indicator of the performance of a Unix-like system; hence, multiple tests are used to test various aspects of the system's performance. These test results are then compared to the scores from a baseline system to produce an index value, which is generally easier to handle than the raw scores. The entire set of index values is then combined to make an overall index for the system. For more information, you can review the project website here: https://github.com/kdlucas/byte-unixbench Each run we spun up an instance with default settings (no optimizations) and the test data we collected is from a full UnixBench test with the iteration count set to 1. Each instance tested generated two entries, one with a single core and another with the maximum number of cores on the instance (up to 36 cores). If the instance only had one core, just one entry was generated. BCU: System Specs Burstorm uses the standard UnixBench score but scaled to a more modern processor and bus (a Raspberry Pi 2, ARM7 @900Mhz) from the original SparcStation 20-61 "George".
  • 38. Rice Burstorm Price Performance Benchmark Report (RB-A) June 2015 38 Appendix 2: What’s next We normalized the values against the score of a Raspberry Pi 2, ARM7 @900Mhz and thus provide a relative score to focus on relative performance more than on absolute performance. This was done because we intend to bring this benchmark data into the Burstorm CAD application to optimize design decisions by performance and price- performance soon. We realize that UnixBench provides a particular test of a UNIX system but it is widely accepted as a measurement for relative performance. We intent to enhance the I/O section because of the potential impact of larger CPU Cache and SSDs on the current tests. As part of RB-B we are considering adding benchmarks for Memory and Network. The first because we see that UnixBench seems only marginally impacted by additional memory while we know certain workloads clearly benefit from memory. The Network aspect is very interesting as it is the most widely shared resource and likely the most volatile. Also it is the most complex to test since by definition a network has dependencies on distance (within the VM, within the OS, within the system, within the local network and so forth and so on). We have plans in progress but welcome contributions from the community. We are also considering how to add more providers to the benchmark. The current Burstorm product catalog has already identified 943 compute & storage cloud services providers. Beyond those we’re working with enterprises to benchmark internally available compute & storage cloud services. The longer term vision for the benchmark framework is to include multi instance benchmarks. Because the Burstorm CAD application is designed to define a complete architecture we see the possibility to then deploy it and run the RB-Benchmark on it to get an overall view of the relative performance of such design. This is obviously a complex goal and will take some time to evolve.
  • 39. Cloud Computing Benchmark 39 Review History We’d like to thank all the reviewers below as well as those who chose to remain anonymous for their contributions to this report. Name Title Affiliation Ravi Anadwali Senior Manager Splunk Darren Bibby VP, Channels & Alliances IDC Mauricio Carreno Senior Manager Accenture Mexico Larry Carvalho Lead Analyst IDC Adrian Cockcroft Technology Fellow Battery Ventures Mac Devine VP, CTO IBM Angel Luis Diaz IT Specialist, Infrastructure IBM Mark Egan Partner Stratafusion Jim Enright Director of Performance Tim Fitzgerald VP Cloud Avnet Sandeep Gopisetty Distinguished Engineer IBM Dave Hansen VP and General Manager Dell Andrew Hately CTO IBM Bill Heil SVP, Chief Bottle Washer VMware Kristopher Johnston Director IT Fidelity Investments Sam Kamal Global Technology Executive Ingram Micro Sunil Kamath Dir. Performance Engineering IBM Ed Laczynski Co-Founder and CEO Zype Cary Landis Solutions Architect SAIC Charles Levine Principal Program Manager Microsoft Dan Ma Assistant Professor Singapore Mngt University William Martroelli Principal Analyst Forrester Research Michael McCain Enterprise Architect Red Hat Justin Mennen VP Enterprise Architecture Estee Lauder Ken Murdoch VP IT & Bldg Operations Save the Children Thao Nguyen Engineer - OCE Facebook Sanjay Rao Associate Professor Purdue University Farhad Shafa Solutions Architect Kaiser Permanente Lloyd Taylor CIO Originate David Wallom Associate Professor University of Oxford Ray Wang Principal Analyst Constellation Research

Related Documents