Multi-Tenant Performance SLOs
Strategies for optimizing multi-tenant performance objectives while managing operating costs efficiently in cloud services. This study delves into tenant scheduling policies, hardware provisioning, and maximizing multi-tenancy to meet service level objectives effectively.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
1 Towards Multi-Tenant Performance SLOs Willis Lang*, Srinath Shankar+, Jignesh M. Patel*, Ajay Kalhan^ *University of Wisconsin-Madison +Microsoft Gray Systems Lab ^Microsoft Corp. To appear in ICDE 2012
Overall Operating Costs of Providing Cloud Services are High 2 Monthly Cost of 46,000 Server Data Center [Hamilton, 2011] Networking Networking $260,039 8% Infrastructure Infrastructure $130,019 4% Servers $1,852,778 Power $1,007,651 Server & Power 88% Dominating costs are server and power costs: 57% and 31% respectively
Performance Service Level Objectives and Managing Cloud Costs 3 Tenants can get their own server and high performance Performance per Tenant Tenants have performance objectives Consolidate tenants onto the fewest number of servers (maximize the degree of multi-tenancy) while maintaining perf objectives Data Center Costs
4 An Optimization Problem Given: Groups of tenants with different performance objectives and a number of server configurations High Perf Low Perf Find: (1) Tenant Scheduling Policies and (2) Hardware Provisioning Policies Such that costs are minimized and performance is delivered
5 Multi-Tenant Scheduling H Tenants L Tenants Perf Objective TPC-C throughput H tenants 100tps L tenants 10tps 1 5 20 1 20 40 #H: #L: Want to maximize degree of multi- tenancy without breaking SLO What if we also have different server types available? Avg H Perf tps ea. 2000 900 130 Avg L Perf tps ea. 2000 110 30
6 Hardware Setup 2 x Intel Nehalem L5630 32GB DDR3 RAID battery-backed cache 1 x 10k RPM SAS OS/software + diskC - $4000 ($111 per month) Data: 2 x 10k RPM SAS 300GB Log: 1 x 10k RPM SAS 300GB ssdC - $4500 ($125 per month) Data: 2 x Crucial C300 256GB Log: 1 x Crucial C300 256GB
7 Software Setup SQL Server 2012 All tenants of the H performance class get an individual database within a SQL Server instance Databases in SQL Server have their own physical files for data and log All tenants of the L performance class get an individual database within a different SQL Server instance SQL Server instance memory provisioning to control performance (not VM)
8 Heterogeneous SLO Characterization diskC 25 25 ssdC 25 Num of H (100tps) Tenants Benchmark server to find max degree multi-tenancy for perf objectives 20 Num of H (100tps) Tenants Num of H (100tps) Tenants 20 20 Systematically reduce H tenants, steadily increase L tenant scheduling until a perf objective fails 15 15 15 10 10 Server characterizing function: 10 Both perf objectives met 5 5 5 Some perf objective fails 0 0 0 0 0 50 50 100 100 0 50 100 Number of L (10tps) Tenants Number of L (10tps) Tenants Number of L (10tps) Tenants
Applying Our Optimization Framework Scenario: 10,000 tenants, 2,000x100tps & 8,000x10tps 9 Optimal Solution: 94 ssdC servers, 38 10tps tenants and 20 100tps tenants + 5 diskC servers, 25 10tps tenants and 20 100tps tenants + 43 ssdC servers, 100 10tps tenants ssdC diskC 40 Number H (100tps) Tenants 30 94 5 20 10 43 0 38 0 50 100 150 Number of L (10tps) Tenants
Applying Our Optimization Framework 10 $30,000 $25,000 Monthly Server Costs $20,000 ssdC 100tps tenants diskC 10tps tenants $15,000 $10,000 $5,000 $0 Optimal Only diskC Tenant Segregated
11 Summary We have presented an optimization framework that tells a Database-as-a-Service provider how to provide performance Service Level Objectives while minimizing cluster infrastructure costs
12 Thesis Research An determine scheduling and server provisioning in light of tenant performance goals [ICDE 2012] server clusters to be much larger and more expensive traditional clusters [DaMoN 2010 Best Paper] Parallel data bottlenecks such bandwith and algorithmic choices are a cause of energy inefficiency [Under Submission] [SIGMOD Record 2009] [CIDR 2009, IEEE DEB 2011] Data Center Costs optimization Complex workloads speedup and force low-power Computational complexity of MR jobs affects the ability to save energy by using smaller clusters [VLDB 2010] By exploiting existing replication schemes, an elegant relationship between load energy efficiency can be exploited parking) and algorithmic choices framework optimal parallel cause to Cluster Design, Performance in the Cloud Low-Power Server Hardware the tenant analytic non-linear Performance ICDE 12 DaMoN 10, Under Submission Characterizing Performance vs Energy and Server Costs than Demonstrated that it is possible to decrease energy and performance in a controlled way using hardware mechanisms frequency/voltage processing as network balancing and (e.g., CPU memory and Cluster-level Performance and Energy Consumption VLDB 10, SIGMOD Rec 09 Node-local Performance and Energy Consumption CIDR 09, IEEE DEB 11
13 Acknowledgements Special thanks to David DeWitt, Jeff Naughton, Alan Halverson, Eric Robinson, Rimma Nehme, Dimitris Tsirogiannis, Nikhil Teletia, Chris R Cluster Design, Performance in the Cloud Low-Power Server Hardware ICDE 12 DaMoN 10, Under Submission Characterizing Performance vs Energy and Server Costs Funded by a grant from Microsoft Gray Systems Lab Cluster-level Performance and Energy Consumption VLDB 10, SIGMOD Rec 09 Node-local Performance and Energy Consumption CIDR 09, IEEE DEB 11
16 Memory-based resource governor E.g., 2 performance goals, 100tps and 10tps 20 tenants pay for 100tps and 30 tenants pay for 10tps The aggregate memory for all 100tps tenants: 20 20 + 30+ 20 100 20 100 + 30 10 2 ??? Similarly, for 10tps tenants: 30 30 10 20 + 30+ 20 100 + 30 10 2 ???
17 Simplicity vs Cost 2.0 diskC cost -10% vs ssdC None of these heuristic methods consistently provides solutions near to the optimal method. 1.5 Rel. Cost 1.0 0.5 Methods ssdC SKU diskC SKU 0.0 20% 100tps, 80% 10tps 50% 100tps, 50% 10tps 80% 100tps, 20% 10tps Hetero SLO Hetero SLO Optimal 2.0 diskC cost -30% vs ssdC Hetero SLO NA ssdC-only 1.5 Rel. Cost NA Hetero SLO diskC-only 1.0 High-perf Low-perf ssdC-H 0.5 0.0 Low-perf High-perf ssdC-L 20% 100tps, 80% 10tps 50% 100tps, 50% 10tps 80% 100tps, 20% 10tps
18 Log Disk Bottlenecks Avg Log Write Wait (ms) Achieved tps for 100tps Tenant 14 180 Average Log Write Wait TPS Achieved by One 160 12 140 100tps Tenant 10 120 Time (ms) 8 100 80 6 60 4 40 2 20 0 0 75/1 100/1 <# 1tps tnt>/<# 100tps tnt> 125/1 150/1 175/1 200/0
