Summary of the content on the page No. 1
White Paper
Consolidation of a Performance-
Intel® Xeon® Processor
7400-based Server
Sensitive Application: Virtualizing
Electronic Sports League’s
Gaming Infrastructure
Abstract
An end-user case study with Electronic Sports League (ESL) using VMware ESX 3.5U1*
with NetQueue feature running on Intel® Xeon® processor 7400 series-based servers with
Intel® NICs supporting VMDq feature. Game servers are mission-critical, single-threaded,
processor-intensive, and network-latency sensitive. Th
Summary of the content on the page No. 2
White Paper Consolidation of a Performance-Sensitive Application Table of Contents Introduction – Can We Virtualize Everything? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 “Non-virtualizable” applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 . Sidebar: ESL .
Summary of the content on the page No. 3
White Paper Consolidation of a Performance-Sensitive Application Introduction — Can We ESL Virtualize Everything? Electronic Sports League (ESL) is the largest online gaming community in Europe, with more than 844,000 active Virtualization of enterprise data center applications using hyper- users as of August 12, 2008. visors or VMMs is taking a predictable path. It started with the consolidation of the simplest, least performance-sensitive, and ESL has deployed thousands of game serve
Summary of the content on the page No. 4
4 sockets with 6 cores Intel Xeon Intel Xeon 4x 7400 Series 1066 Intel Xeon Intel Xeon Intel Xeon processor 7300 series ESB2 I/O Bridge Configurable PCI Express* White Paper Consolidation of a Performance-Sensitive Application Problem statement ESL’s perception, along with that of many of its peers in the gaming industry, was that gaming servers “can’t be virtualized” due to their mission-critical nature, high CPU utilization, and the critical focus of their user base on the in-game tran
Summary of the content on the page No. 5
VM VM VM n 1 2 Virtualization Hypervisor NIC White Paper Consolidation of a Performance-Sensitive Application Virtual Virtual Virtual 10.0 8.0 Unused I/O capacity 6.0 4.0 4.0 2.0 without VMDq LAN Result: NIC performance can be up to ~60% underutilized Figure 3 . Network data flow for virtualization without the use of Figure 4 . Impact of virtualization on a 10 GB Ethernet NIC VMDq and NetQueue technologies . without the use of VMDq and NetQueue . of virtualization in the Intel lab using
Summary of the content on the page No. 6
VM VM VM 1 2 n VMware with NetQueue NIC with VMDq White Paper Consolidation of a Performance-Sensitive Application Virtual Virtual Virtual 10.0 9.5 9.2 8.0 6.0 Without VMDq With VMDq 4.0 With VMDq 4.0 Jumbo Frames 2.0 LAN • 2x throughput • Near Native 10 GbE Figure 5 . Network data flow for virtualization Figure 6 . Tests measure wire speed Receive (Rx) side performance with VMDq and NetQueue . with VMDq on Intel® 82598 10 Gigabit Ethernet Controllers . VMM overhead can be directly routed
Summary of the content on the page No. 7
Counter- Counter- Strike 1.6* Strike 1.6* Windows Windows Server 2003* Server 2003* 32 bit 32 bit VM1 VMn White Paper Consolidation of a Performance-Sensitive Application 25 20 15 10 5 VMware ESX 3.5 U1* Best LAN Best Typical Intel® Xeon® Processor 7400 Series-based Server Internet Internet Figure 7 . Software stack . Figure 8 . In-game latency . Software stack Question the Assumptions Figure 7 shows the software stack used in the PoC. We started with VMware ESX 3.5U1, which was the late
Summary of the content on the page No. 8
Client 1 Client 8 White Paper Consolidation of a Performance-Sensitive Application 1 GbE 40–200 bytes UDP 10 GbE Native Force 10 S50* Game Player Game Server Server 2,000 bytes UDP 1 GbE Figure 9 . Game network protocol overview . Figure 10 . Native lab test setup . Gaming protocol Round-trip network latency tests How does this gaming protocol work? A player sends a UDP In order to run the round-trip network latency test, we used the packet of 40–200 bytes in size to the game server. The g
Summary of the content on the page No. 9
Client 1 VM1 vSW1 VM8 Client 8 White Paper Consolidation of a Performance-Sensitive Application 0.30 0.25 0.20 0.15 0.10 ESX* Server 1 GbE 0.05 10 GbE Force 10 S50* 64 256 1024 Packet size (bytes) 1 GbE Native VMDq No VMDq Figure 11 . Virtualized lab test setup . Figure 12 . Netperf 2 .4 .4 UDP latency test with eight parallel streams . Scenarios 2 and 3: Virtualized with The results from these tests are summarized in Figure 12. In and without VMDq/NetQueue the graph, we are comparing n
Summary of the content on the page No. 10
VM1 VMn Svc Cons vSwitch 1 vSwitch 0 Force 10 S50N Counter- Counter- ESL Network Strike 1.6* Strike 1.6* Windows Windows Server 2003* Server 2003* 32 bit 32 bit VMn VM1 White Paper Consolidation of a Performance-Sensitive Application 10 GbE VMDq 1 GbE Firewall VMware ESX 3.5 U1* Internet Intel® Xeon® Processor 7400 Series-based Server Figure 13 . Live Internet test setup . Figure 14 . Test software stack . 2 GB memory, Windows 2003 server (32 bit) and Counter-Strike ESL game testing 1
Summary of the content on the page No. 11
White Paper Consolidation of a Performance-Sensitive Application 100 740 720 80 700 60 680 40 660 20 640 24 32 36 40 Number of VMs Figure 15 . ESL virtual game servers on Intel® Xeon® processor 7400 Series-based platforms . latency or gaming experience. But in ESL’s current native envi- ROI Delivered ronment, ESL tries to limit the CPU utilization per server around 60–80 percent. ESL needs this head room as there can be situa- Server consolidation tions where some game servers need more r
Summary of the content on the page No. 12
White Paper Consolidation of a Performance-Sensitive Application Table 1 . ESL standard native hardware deployment metrics (“Before”) vs . virtualized deployment Proof-of-Concept metrics as described in this paper (“After”) . Before After Server 1P Intel® Core™2 Duo processor 4P Intel® Xeon® processor 7400 Cores 2 24 Game server processes 3 per CPU core 3 per VM; 4.5 per CPU core VMs per box N/A 36 Game server per box 6 108 Total users per box 72 1,296 CPU utilization 60-80% 75% Power uti
Summary of the content on the page No. 13
White Paper Consolidation of a Performance-Sensitive Application Conclusions We have shown that ESL’s mission-critical, processor-intensive, “T he new Six-Core Intel Xeon 7400 processor series network latency-sensitive game servers can be very effectively (“Dunnington”) was completely overwhelming in virtualized using the latest technologies. We achieved 18:1 all terms . The Intel Xeon MP servers with Intel consolidation ratios and saved almost 90 percent on power VMDq technology en
Summary of the content on the page No. 14
White Paper Consolidation of a Performance-Sensitive Application Resources Additional sources of information on this topic Intel Virtualization Developer Community: www.microsoft.com/windowsserver/compare/compare-windows-to-unix.mspx Intel Virtualization Developer Community: softwarecommunity.intel.com/articles/eng/1424.htm Intel Virtualization Technology: www.intel.com/technology/virtualization/index.htm Intel Xeon processor 7400 series: www.intel.com/performance/server/xeon_mp/virt
Summary of the content on the page No. 15
Summary of the content on the page No. 16
We want to hear from you Join the conversation on Intel’s Open Port: The Server Room • Bring your questions to the “Ask an Expert” discussion forum • Share best practices with IT influencers • Engage with Intel on your Datacenter challenges P e r fo r ma n c e t e s t s a nd r a t i n gs a re m ea s u re d u s i ng s p ec i fi c c o mpu t e r s y s t e ms a nd / o r c o mp o n e n t s a nd re fl ec t t he a ppr o x i ma t e p e r fo r ma n c e o f I n t e
