1 Realization of the Low Cost and High Performance MySQL Cloud Database Wei Cao Alibaba Cloud Computing Ltd. No. 969,Wenyi Road, Hangzhou, Zhejiang Province,China Feng Yu Alibaba Cloud Computing Inc. No. 969,Wenyi Road,Hangzhou, Zhejiang Province,China Jiasen Xie Alibaba Cloud Computing Inc. No. 969,Wenyi Road,Hangzhou, Zhejiang Province,China ABSTRACT MySQL is a low cost, high performance, good reliability and open source database product, widely used in many Internet companies. For example, there are thousands of MySQL servers being used in Taobao. Although NoSQL developed very quickly in past two years, and new products emerged in endlessly, but in the actual business application of NoSQL, the requirements to developers are relatively high. Moreover, MySQL has many more mature middleware, maintenance tools and a benign ecological circle, so from this perspective, MySQL dominates in the whole situation, while NoSQL is as a supplement. We (the core system database team of Taobao) have done a lot of work in the filed of MySQL hosting platform, designed and implemented a UMP (Unified MySQL Platform) system, to provide a low cost and high performance MySQL cloud database service. 1. INTRODUCTION UMP (Unified MySQL Platform) system is a low cost and high performance MySQL cloud scheme, which is developed by the core database team of Taobao. It's key module is realized by the Erlang programming language. The system including controller server, proxy server, agent server, API/Web server, analysis and optimization server and log analysis server is depended on the Mnesia, LVS, RabbitMQ, ZooKeeper and other open source components. Developers can apply for MySQL instance resources from the platform, and access the data through a single entrance provided by the platform. UMP system maintains and managements a resource pool, to provide master-slave hot standby, data backup, disaster recovery, R/W splitting, database sharding and other service to user transparently. UMP System reduces the overall cost by a variety of resource virtualization methods, such as, sharing one MySQL instance for multiple small users, letting medium-sized users have a MySQL This work is licensed under the Creative Commons Attribution- NonCommercial-NoDerivs 3.0 Unported License. To view a copy of this license, visit Obtain permission prior to any use beyond those covered by the license. Contact copyright holder by ing Articles from this volume were invited to present their results at the 40th International Conference on Very Large Data Bases, September 1st - 5th 2014, Hangzhou, China. Proceedings of the VLDB Endowment, Vol. 7, No. 13 Copyright 2014 VLDB Endowment /14/08 instance themselves and multiple MySQL instances sharing the same physical machine. In the aspect of resource isolation, UMP system realizes resource virtualization and guarantees the quality of service for users by restricting MySQL process resources through the Cgroup, and limiting QPS in proxy server. In addition, with the integrated use of SSL, IP white list, user log and SQL intercept, UMP system can protect the security of user data. This paper is organized as follows. Section 2 introduces the overall design of UMP system, the session 3 depicts the system interactions, the session 4 discusses UMP s high availability, the session 5 tells the implementation of the key system function, the session 6 summarizes the whole text. 2. DESIGH OVERVIEW To meet the needs of most users, UMP system is designed to serve users transparently, that means users are not aware of perceiving the events of downtime, disaster recovery, flash disconnect and R/W splitting. Our goal is to provide a unified interface, which shields the underlying implementation land provides high availability of services. 2.1 Architecture The architecture of the first edition of UMP system is based on mysql-proxy 0.8. We have fixed several bugs of mysql-proxy, and made some modifications on the state machine which deals with user connections and database connections in the proxy plug-in. We also developed Lua script to implement some functions, such as obtaining the user authentication information and the address of back-end database from a central database, establishing a connection to the backend database, forwarding packets, and etc. However, in the process of developing and deploying the first edition, we realized a few problems. First, mysql-proxy 0.8 only provides simple and crude supports on multithreading where multiple threads share the same message queue and listen to the same socket pair channel, which often causes "Thundering Herd" phenomenon. In addition, mysql-proxy uses global Lua lock, which means only one worker thread can execute Lua scripts (although improvements will be made in version 0.9). As a result, the performance of mysql-proxy under multi-threaded mode is far from keeping linear growth to the number of CPU cores. Second, the framework of mysql-proxy limits our extension for UMP system, such as achieving the user's connection limit, QPS 1742
2 LVS Proxy Server Data Route Resource Control Database Sharding Flow Control Log Analysis Dual Master Agent Server Task Execution Instance Control Monitor MySQL Cluster User Controller Server Metadata Cluster Management Instance Management Message Queue Disaster Recovery Data Backup Database Expansion ZooKeeper High Availability Member Management UMP Cluster Basic Services RabbitMQ Communication API / Web Server Configuration Portal State Query Log Analysis Server Expert System Information Statistics limit, master-slave switch, R/W splitting, database sharding, and a series of other functions. As a result, we decided to rewrite a proxy server to replace the mysql-proxy by using the Erlang programming language. Figure 1 shows the architecture of the current UMP system. The system includes controller server, proxy server, agent server, API/Web server, log analysis server and information statistical server. Also some open source components are involved, such as LVS, RabbitMQ, ZooKeeper and so on. The entire project currently has 50,000 lines' Erlang code, lines' C/C++ source code, and lines' source code in other languages. Erlang is a structured, dynamic and functional programming language. OTP is a framework and platform for the development of distributed and high fault-tolerant Erlang applications. The reason why we used Erlang language to rewrite the proxy server was that Erlang/OTP perfectly abstracts the required elements for developing a distributed, highly fault-tolerant application, including network programming framework, serialization and deserialization, fault tolerance and hot deployment. 2.2 Interface We developed a single unified entrance to meet the user s needs which performing all operations through webpages, including create and delete instances, DB control, promotion and demotion, transfer, and etc. The API/Web server provides the system management interface for users, which is based on open source projects called Chicago Boss and Mochiweb. Mochiweb provides the http/https service, while Chicago provides MVC frameworks like Rails Chicago. Figure 1. Architecture diagram of UMP system. Chicago has good support for concurrency, each request of it occupies a lightweight Erlang process. In addition, the interfaces we designed is compatible with the interfaces of AWS. 2.3 Proxy The first important reason why we developed the proxy is to fix the defects from mysql-proxy, and the other one is to realize a transparent service mechanism. At the same time, we can add some features on proxy, such as flow control and log analysis. The proxy server provides the service to access the MySQL database for users, which fully implements the MySQL protocol. The users can use the existing MySQL client to connect to the proxy server, and then the proxy server gets the authentication information of users, resources quotas limit (for example the maximum number of connections, QPS, IOPS, etc.) by user name, as well as the address of MySQL instance (list), after that it transmits the user's SQL queries to the correct MySQL instance. In addition to the basic functions of data routing, the proxy server also implements the following functions, including limiting resource, shielding MySQL instance faults, R/W splitting, database sharding, recording user access logs, and etc. The proxy servers are stateless in order to provide high availability service. 2.4 Metadata We store the metadata which is used by the UMP system in the metabase of the controller server. The stored metadata is mainly including cluster membership information, user configuration and status information, the mapping between MySQL instances, task 1743
3 information, and backup migration information. In order to achieve distributed storage of metadata, we put the operations of reading and writing metadata into a set of Mnesia distributed database services. Other server components can obtain the corresponding data only by sending a request to the controller. 2.5 Controller UMP needs a perfect management and control service for the backend cluster, so our controller server provides a variety of management services for UMP cluster, including metadata storage, cluster membership management, MySQL instance management, backup, migration, expansion and other functions. The cluster membership management, including cluster management division, host distribution, logical group configuration, and etc. is used to carry out overall control of instances, such as upgrade and migration. The management operations of a MySQL instance, like creation, deletion, migration and expansion, are all encapsulated into a task respectively, these tasks will be broken down into idempotent sub-steps according to business needs, that means even the same steps are executed repeatedly by the agent server, there are not any side effects, so it offers retry operation when the task fails. The agent server is deployed on a machine running MySQL processes, used to manage MySQL instances on the physical machine, performing the actual operation, such as create, delete, backup, migration and master-slave switching. 2.6 Communication The communication between nodes in UMP system (excluding the transfer of large data in SQL queries, logs and other streams, these go directly to these TCP's) are all through RabbitMQ, which is used as a messaging middleware to ensure the reliability of the message communication. When initialized, the cluster will create a queue for each node in the cluster in RabbitMQ, which is used as a mailbox for each node. When a node want to send a message to other node, it just write a message to the "mailbox" of the other side, regardless of whether the other node is online or not, then the RabbitMQ client running on the other side will receives the message,and call the appropriate handling routine. After the message is processed, the client will return an ACK packet to RabbitMQ, and the RabbitMQ will delete this message from the "mailbox". We can achieve RPC based on RabbitMQ, the client also writes a message to the sender's Reply "mailbox" in addition to send a ACK packet to RabbitMQ to delete the Request message. RabbitMQ is transaction supported, it can guarantee delete Request and reply message are done in an atomic operation, hereby providing a reliable communication service. 3. SYSTEM INTERACTIONS Our design should simplify all operation since the UMP system contains multiple components. As a result, we created this cooperation job system with multiple components to help users to do with instance controlling, disaster recovery, dilatation and other functions. 3.1 Control Order The controlling operation designed in UMP system is launched by API/Web servers, processed by a series of servers and finally sent to the machine holding MySQL instance to do the work. Task processing and scheduling is transparent to users and the unitized interface brings a convenient experience to users. In Figure 2, we illustrate this process by following the control flow of instance-create through these numbered steps. 1. The clients launch a creating-instance task by accessing the API/Web server. 2. When the controller server receives the notice from API/Web server, it initializes the context of working flow and assigns the steps of the task to memory. 3. The steps to be executed will be distributed to corresponding agent server according to business. The agent server runs the steps and creates a MySQL instance in physical machine. 4. The progress and result that the agent server ran the task will all be fed back to controller server. 5. As soon as the feedback is received by controller server, it will record the behavior and result of the execution to meta database. So will these information be displayed on Web Console UI. Client LVS Proxy MySQL Agent 1 API\WEB Log Analysis Expect System 3 4 Controller Others controlling processes are similar as above, and controller server can do work concurrently between different instances. 3.2 Data Flow Data, which refers to the DML statements or Query statements that manipulate the database here, is separated with control flow in system. This will achieve better controlling and management. Data chain flows through LVS first, converting user's link string to VIP and VPORT, which help to build connection to the server. After the data flow passes LVS and reaches proxy server, the router function will do its work and transfer user's SQL statement to the right SQL instance. The data will be recorded in user access log while passing proxy server which will be processed after fetching by the log analysis server. 3.3 Operation and Maintenance In order to better understand cloud database running status, UMP party provides an operational log system, mainly is the collection 2 5 Legend: Figure 2. Control order and data flow. Control Data 1744
4 and analysis of relevant logs and presented to the system management interface, including log analysis, expert system, etc Portal Considering the interaction of unity, the portal is present in the system management interface of API/Web server. The user s access information is fetched by log analysis server and the optimization information provided by expert system will be sent to the Portal on the rendering. The information statistics server will show the connection numbers, QPS numerical and MySQL instance status on the Portal, for the future implementation of elastic resource allocation and the automated MySQL instance migration provides gist Log Analysis For the convenience of users to analysis their operation of SQL statements, log analysis server will store and analyze the users access logs incoming by proxy server, for the execute slower logs abstract the sentence mode,and implement real-time indexes for user query over a period of slow logs and statistics Expert System Often some users want to know the slow logs which impact their business performance, our expert system will provide the SQL statement analysis and optimization services not only collect the slow logs of MySQL instance for users,but also integrates and analysis these logs, refining to every slow logging statements, analyze the slow reason, and provide our optimization scheme. 4. HIGH AVAILABILITY As described above, UMP system has many components. The high availability of some key components among them is particularly important. Also their quality of service determines the overall performance of the system and the impact on user. 4.1 Dual Master Replication For each user system, UMP maintains master and slave MySQL instance with structure of dual master. The proxy will guarantee to write data to master instance. When the database fails, controller server will detect this event and launch a master-slave switching operation, which will inform all proxy server to perform the switching through the RabbitMQ. When the fault master instance online again, controller server will stop writing slave instance until master and slave databases are consistent and inform the proxy server to switch the write operation to the master instance. This process will make users feel cannot write for a short period of time. System performs a transparent recovery process to users, so users cannot perceive the master instance downtime and launch event. The proxy provides a persistent connection for user in order to reduce the times of flash disconnection to some extent. 4.2 Controller Leader In order to achieve high availability, the system will deploy multiple controller servers. They will elect a leader by the distributed lock algorithm provided by the ZooKeeper, which is responsible for scheduling and monitoring various system tasks, such as create or drop instance, backup and migration. These tasks would be divided into multiple steps, and would involve multiple components in the system, such as the main library, the library and the proxy server. In order to provide the function of rollback, we adopt a similar workflow way to achieve it. Each system tasks is divided into Erlang process of multiple stages, after each step and before executing the next step, the middle state is persisted to the Mnesia. If the task is stopped because of a node failure, the leader will detect and re-initiate the task that will go on from the last failed "breakpoint". 4.3 Stateless Proxy The proxy servers are stateless, so proxy server downtime will not affect the other servers in the system, which will only disconnect a user from the proxy server. Multiple proxy servers using LVS HA program to achieve load balancing, user application will be redirected to other LVS's proxy after re-connecting, so as to provide high availability level expansion. 5. IMPLEMENTATION Platform reduces cost through running multiple MySQL instances on a single physical machine, and realizes resource isolation, distribution according to needs and limitations of CPU, memory and IO resources control. The inner maintenance and resource pool management of UMP system provides a series of service such as master-slave hot standby, data backup, migration, R/W splitting, database sharding and data security transparently. 5.1 Resource Control Referring to the resource management method in some cloud computing system such as the VMware DRS, we implemented a resource pool mechanism to manage the database on the server's CPU, memory, disk and other computing resources. In addition to not affect the user experience, we realized the resource isolation mechanism for MySQL instance Resource Allocation Allocated instance is the unit of resource pool, the administrator can deploy in what rooms, according to the application need to be separately specified computing resources of the library, from the library's resources pool, instance management services choose the lighter server from the resource pools to create an instance of server. Moreover, combined with Cgroup will further resources to facilitate the management and isolation, at the same time it can limit the upper limit of each process group. Now the system supports three kinds of specifications of the users: small flow users that share the same MySQL instance, mediumsized users that possesses a MySQL instance alone and possession of multiple independent MySQL instance depots table users. For different users, we provide different kinds of service of dynamic expansion and contraction Resource Isolation Isolating resource is particularly important when multiple users share one MySQL instance or multiple MySQL instances share the same host. Currently, we implement the function of resource isolation through combing the use of Cgroup on the database servers to limit MySQL instances resource and the limiting of QPS on proxy servers. 1745
5 Firstly, we use the cpuset, memcg and blkio sub-modules of Cgroup to limit the maximum CPU usage, memory and IOPS that would be used by MySQL process. Secondly, we increase the delay on proxy server to limit the user QPS and reduce the system resources consumed. 5.2 R/W Splitting We also realized read/write splitting transparent to users. proxy server will analyze the incoming user SQL statements, will write sent to the master library, read operation load balance distribution to master and slave library. In order to avoid reading before the consistent master-slave synchronization, any data within 300 milliseconds after each write operation to read will be forced to distribute to the master library. Through master-slave multithreaded replication technology, 300 milliseconds can basically guarantee the data from the master library synchronous to slave library, and this value can be regulated in the configuration. Besides, in order to share the pressure of reading and writing, system realized the read-only instance, users can put read-only data into read-only instance, so as to reduce the pressure of the master library. 5.3 Database Sharding The user can specify the user type as multi-instances based on the requirements. This multi-instances type users need to add the rules of database sharding into SQL comments when they are creating tables. The implementation procedure of the database sharding is shown in Figure 3. Firstly, the proxy server will parse the user incoming SQL statements to extract the information which is needed to rewrite and distribute the SQL statements. SQL Parse MySQL Rewrite Distribute Queries MySQL Protocl Logic Control Layer MySQL Protocl Merge Sort Buffer Control Figure 3. Process of database sharding. Secondly, the SQL statement is rewritten in the form of substatement to be executed on each of the separated table. This mainly include replacing the table names and where conditions. And then the sub-statements are sent to each sub-table to be executed concurrently. Finally, the result returned from each sub-table is received and merged. To avoid the size of query result set becoming too large, we limit the number of results rows of each MySQL instance by setting the buffer size. At the moment all the sub-tables returned some results the merge sort begin to be performed, and then the merged result is returned to the user. In order to improving the performance, we used C++ to achieve the parsing and rewriting of the SQL statement and the merging of the results returned from MySQL servers, which is called by the Erlang language state machine through the NIF interface. 5.4 Data Security User and enterprise security department will be more concerned about the security of their data, we implemented a variety of methods to ensure the safety of the user data: 1. UMP supports the SSL protocol. The proxy server realized the whole MySQL client/server protocol, which can establish an encrypted connection with the client. 2. System generates a list of IP address which is used to access to database by setting up white list. Our users can configure their application server address into the white list in order to increase the security of the account. 3. The proxy server will put all user database operation record to log analysis server, so security department can export log text on a regular basis and scan security vulnerabilities. 4. The proxy server can intercept various types of SQL statements according to the requirement of the security department, such as select * statement on a full table, statements whose result is beyond limit, etc. 6. CONCLUSIONS Some components in UMP system, such as the proxy server and the server log analysis, has been used in the Poly Tower Platform of Tmall to provide secure data cloud services for electricity and ISV. In addition, the UMP system is also used in the decorating platform of Taobao shop to provide data services for developers. In the next stage, we hope the UMP system can further reduce the cost of internal data storage. In engineering practice, we adhere to the principle of not reinventing the wheel through taking advantage of open source, mature technology and tools. For example, we implemented the high-performance proxy server on the Erlang network programming framework and the messaging middleware based on RabbitMQ, we manage the server heartbeat using ZooKeeper, and we also make full use of mature tools in the internal group, such as data backup, migration, programs of capacity expansion or contraction and other bin log tools. This principle allows us to concentrate on reducing costs and improving user experience with the limited resources. 7. REFERENCES  Mattsson H, Nilsson H, Wikström C. Mnesia A distributed robust DBMS for telecommunications applications[m]//practical Aspects of Declarative Languages. Springer Berlin Heidelberg, 1998:
6  Kramer J. Advanced message queuing protocol (AMQP)[J]. Linux Journal, 2009, 2009(187): 3.  Hunt P, Konar M, Junqueira F P, et al. ZooKeeper: Wait-free Coordination for Internet-scale Systems[C]//USENIX Annual Technical Conference,2010, 8:
Cloud Based Application Architectures using Smart Computing How to Use this Guide Joyent Smart Technology represents a sophisticated evolution in cloud computing infrastructure. Most cloud computing products
High Availability Solutions for the MariaDB and MySQL Database 1 Introduction This paper introduces recommendations and some of the solutions used to create an availability or high availability environment
Module 14: Scalability and High Availability Overview Key high availability features available in Oracle and SQL Server Key scalability features available in Oracle and SQL Server High Availability High
MyISAM Default Storage Engine before MySQL 5.5 Table level locking Small footprint on disk Read Only during backups GIS and FTS indexing Copyright 2014, Oracle and/or its affiliates. All rights reserved.
Eloquence Training What s new in Eloquence B.08.00 2010 Marxmeier Software AG Rev:100727 Overview Released December 2008 Supported until November 2013 Supports 32-bit and 64-bit platforms HP-UX Itanium
ORACLE DATABASE 10G ENTERPRISE EDITION OVERVIEW Oracle Database 10g Enterprise Edition is ideal for enterprises that ENTERPRISE EDITION For enterprises of any size For databases up to 8 Exabytes in size.
Network Attached Storage Jinfeng Yang Oct/19/2015 Outline Part A 1. What is the Network Attached Storage (NAS)? 2. What are the applications of NAS? 3. The benefits of NAS. 4. NAS s performance (Reliability
Cisco Application Networking Manager Version 2.0 Cisco Application Networking Manager (ANM) software enables centralized configuration, operations, and monitoring of Cisco data center networking equipment
Neverfail for Windows Applications June 2010 Neverfail, from Neverfail Ltd. (www.neverfailgroup.com), ensures continuity of user services provided by Microsoft Windows applications via data replication
SCHOONER WHITE PAPER Top 10 Reasons why MySQL Experts Switch to SchoonerSQL - Solving the common problems users face with MySQL About Schooner Information Technology Schooner Information Technology provides
BlackBerry Enterprise Service 10 Secure Work Space for ios and Android Version: 10.1.1 Security Note Published: 2013-06-21 SWD-20130621110651069 Contents 1 About this guide...4 2 What is BlackBerry Enterprise
5th International Conference on Advanced Materials and Computer Science (ICAMCS 2016) Memory Database Application in the Processing of Huge Amounts of Data Daqiang Xiao 1, Qi Qian 2, Jianhua Yang 3, Guang
Volume 1, Issue 1 ISSN: 2320-5288 International Journal of Engineering Technology & Management Research Journal homepage: www.ijetmr.org Analysis and Research of Cloud Computing System to Comparison of
Real-time Data Replication from Oracle to other databases using DataCurrents WHITEPAPER Contents Data Replication Concepts... 2 Real time Data Replication... 3 Heterogeneous Data Replication... 4 Different
Architectures Haute-Dispo Joffrey MICHAÏE Consultant MySQL 04.20111 High Availability with MySQL Higher Availability Shared nothing distributed cluster with MySQL Cluster Storage snapshots for disaster
RingStor User Manual Version 2.1 Last Update on September 17th, 2015 RingStor, Inc. 197 Route 18 South, Ste 3000 East Brunswick, NJ 08816 Page 1 Table of Contents 1 Overview... 5 1.1 RingStor Data Protection...
DISTRIBUTED SYSTEMS [COMP9243] Lecture 9a: Cloud Computing Slide 1 Slide 3 A style of computing in which dynamically scalable and often virtualized resources are provided as a service over the Internet.
4th International Conference on Mechatronics, Materials, Chemistry and Computer Engineering (ICMMCCE 2015) Design and Implementation of IaaS platform based on tool migration Wei Ding State Key Laboratory
WHITE PAPER Affordable, Scalable, Reliable OLTP in a Cloud and Big Data World: IBM DB2 purescale Sponsored by: IBM Carl W. Olofson December 2014 IN THIS WHITE PAPER This white paper discusses the concept
VMware vsphere: Install, Configure, Manage [V5.0] Gain hands-on experience using VMware ESXi 5.0 and vcenter Server 5.0. In this hands-on, VMware -authorized course based on ESXi 5.0 and vcenter Server
Big Data Storage Architecture Design in Cloud Computing Xuebin Chen 1, Shi Wang 1( ), Yanyan Dong 1, and Xu Wang 2 1 College of Science, North China University of Science and Technology, Tangshan, Hebei,
Using MySQL for Big Data Advantage Integrate for Insight Sastry Vedantam firstname.lastname@example.org Agenda The rise of Big Data & Hadoop MySQL in the Big Data Lifecycle MySQL Solutions for Big Data Q&A
Bigdata High Availability (HA) Architecture Introduction This whitepaper describes an HA architecture based on a shared nothing design. Each node uses commodity hardware and has its own local resources
Send Orders for Reprints to email@example.com 1582 The Open Cybernetics & Systemics Journal, 2015, 9, 1582-1586 Open Access The Construction of Seismic and Geological Studies' Cloud Platform Using
Adding Indirection Enhances Functionality The Story Of A Proxy Mark Riddoch & Massimiliano Pinto Introductions Mark Riddoch Staff Engineer, VMware Formally Chief Architect, MariaDB Corporation Massimiliano
Oracle WebLogic Server 11g Administration This course is designed to provide instruction and hands-on practice in installing and configuring Oracle WebLogic Server 11g. These tasks include starting and
ZEN LOAD BALANCER EE v3.04 DATASHEET The Load Balancing made easy OVERVIEW The global communication and the continuous growth of services provided through the Internet or local infrastructure require to
Database as a Service (DaaS) Version 1.02 Table of Contents Database as a Service (DaaS) Overview... 4 Database as a Service (DaaS) Benefit... 4 Feature Description... 4 Database Types / Supported Versions...
In Memory Accelerator for MongoDB Yakov Zhdanov, Director R&D GridGain Systems GridGain: In Memory Computing Leader 5 years in production 100s of customers & users Starts every 10 secs worldwide Over 15,000,000
Virtual Machine in Data Center Switches Huawei Virtual System Contents 1 Introduction... 3 2 VS: From the Aspect of Virtualization Technology... 3 3 VS: From the Aspect of Market Driving... 4 4 VS: From
Ultra-Scalable Storage Provides Low Cost Virtualization Solutions Flexible IP NAS/iSCSI System Addresses Current Storage Needs While Offering Future Expansion According to Whatis.com, storage virtualization
Tier Architectures Kathleen Durant CS 3200 1 Supporting Architectures for DBMS Over the years there have been many different hardware configurations to support database systems Some are outdated others
ZEN LOAD BALANCER EE v3.02 DATASHEET The Load Balancing made easy OVERVIEW The global communication and the continuous growth of services provided through the Internet or local infrastructure require to
the Availability Digest Redundant Load Balancing for High Availability July 2013 A large data center can comprise hundreds or thousands of servers. These servers must not only be interconnected, but they
Chapter 2 TOPOLOGY SELECTION SYS-ED/ Computer Education Techniques, Inc. Objectives You will learn: Topology selection criteria. Perform a comparison of topology selection criteria. WebSphere component
IBM Tivoli Composite Application Manager for Microsoft Applications: Microsoft Hyper-V Server Agent Version 6.3.1 Fix Pack 2 Reference IBM Tivoli Composite Application Manager for Microsoft Applications:
Integrated Application and Data Protection NEC ExpressCluster White Paper Introduction Critical business processes and operations depend on real-time access to IT systems that consist of applications and
Course 10231B: Designing a Microsoft SharePoint 2010 Infrastructure OVERVIEW About this Course This 5 day ILT course teaches IT Professionals to design and deploy Microsoft SharePoint 2010. Audience Profile
MySQL Database for High Availability Web Applications Tushar Joshi Turtle Networks Ltd www.turtle.net Overview What is High Availability? Web/Network Architecture Applications MySQL Replication MySQL Clustering
Features Comparison: Hyper-V Server and Hyper-V February 2012 The information contained in this document relates to a pre-release product which may be substantially modified before it is commercially released.
Adding High Availability to the Cloud Paul J. Holenstein Executive Vice President Gravic, Inc. Companies are moving more and more IT services and utility applications to public clouds to take advantage
Data Management in the Cloud Ryan Stern firstname.lastname@example.org : Advanced Topics in Distributed Systems Department of Computer Science Colorado State University Outline Today Microsoft Cloud SQL Server
Database Resilience at ISPs High-Availability White Paper Internet Service Providers (ISPs) generally do their job very well. The commercial hosting market is segmented in a number of different ways but
Jitterbit allows you to easily integrate Microsoft Dynamics CRM with any cloud, mobile or on premise application. Jitterbit s intuitive Studio delivers the easiest way of designing and running modern integrations
October 2013 Daitan White Paper Reference Model for Cloud Applications CONSIDERATIONS FOR SW VENDORS BUILDING A SAAS SOLUTION Highly Reliable Software Development Services http://www.daitangroup.com Cloud
CMB 207 1I Citrix XenApp and XenDesktop Fast Track This fast paced course provides the foundation necessary for students to effectively centralize and manage desktops and applications in the datacenter
CHAPTER 1 - JAVA EE OVERVIEW FOR ADMINISTRATORS Java EE Components Java EE Vendor Specifications Containers Java EE Blueprint Services JDBC Data Sources Java Naming and Directory Interface Java Message
Client Overview Our client provides leading video platform for enterprise HD video conferencing and has product suite focused on product-based visual communication solutions. Our client leverages its solutions
Migrating to vcloud Automation Center 6.1 vcloud Automation Center 6.1 This document supports the version of each product listed and supports all subsequent versions until the document is replaced by a
Citrix Application Streaming Universal Application Packaging and Delivery Breaking Away from Traditional IT Application Packaging Application Delivery to Endpoint Devices Application Isolation Mode Automatic
Introduction to Hadoop New York Oracle User Group Vikas Sawhney GENERAL AGENDA Driving Factors behind BIG-DATA NOSQL Database 2014 Database Landscape Hadoop Architecture Map/Reduce Hadoop Eco-system Hadoop
VMware vsphere 4.1 Pricing, Packaging and Licensing Overview E f f e c t i v e A u g u s t 1, 2 0 1 0 W H I T E P A P E R Table of Contents Executive Summary...................................................
Database Systems Journal vol. II, no. 2/2011 23 A Grid Architecture for Manufacturing Database System Laurentiu CIOVICĂ, Constantin Daniel AVRAM Economic Informatics Department, Academy of Economic Studies
CA ARCserve Replication and CA ARCserve High Availability r16 CA ARCserve Replication and CA ARCserve High Availability Deployment Options for Microsoft Hyper-V Server TYPICALLY, IT COST REDUCTION INITIATIVES
VERITAS Cluster Server Traffic Director Option Product Overview V E R I T A S W H I T E P A P E R Table of Contents Traffic Director Option for VERITAS Cluster Server Overview.............................................1
Connect With Confidence Astaro Deployment Guide Clustering and Hot Standby Table of Contents Introduction... 2 Active/Passive HA (Hot Standby)... 2 Active/Active HA (Cluster)... 2 Astaro s HA Act as One...
Microsoft SQL Server 2008 R2 Enterprise Edition and Microsoft SharePoint Server 2010 Better Together Writer: Bill Baer, Technical Product Manager, SharePoint Product Group Technical Reviewers: Steve Peschka,
Aug/12/2002 Page 1 of 5 Administering a Microsoft SQL Server 2000 Database Catalog No: RS-MOC2072 MOC Course Number: 2072 5 days Tuition: $2,070 Introduction This course provides students with the knowledge
scalability OneBridge Mobile Groupware technical brief An important consideration for IT professionals when choosing a server-based synchronization solution is that of scalability. Scalability refers to
Application Note Achieving Zero Downtime and Accelerating Performance for WordPress Executive Summary WordPress is the world s most popular open source website content management system (CMS). As usage
Parallels Virtuozzo Containers 4.0 for Linux Release Notes Copyright 1999-2011 by Parallels Holdings, Ltd. All rights reserved. This document describes the new features of this release and important changes
Monitoring DoubleTake Availability eg Enterprise v6 Restricted Rights Legend The information contained in this document is confidential and subject to change without notice. No part of this document may
ACHIEVING 100% UPTIME WITH A CLOUD-BASED CONTACT CENTER Content: Introduction What is Redundancy? Defining a Hosted Contact Center V-TAG Distribution Levels of Redundancy Conclusion Fault Tolerance Scalability
How to Backup and Restore a VM using Veeam Table of Contents Introduction... 3 Assumptions... 3 Add ESXi Server... 4 Backup a VM... 6 Restore Full VM... 12 Appendix A: Install Veeam Backup & Replication
VMware vcloud Automation Center 6.1 Reference Architecture T E C H N I C A L W H I T E P A P E R Table of Contents Overview... 4 What s New... 4 Initial Deployment Recommendations... 4 General Recommendations...
EMC AVAMAR INTEGRATION WITH EMC DATA DOMAIN SYSTEMS A Detailed Review ABSTRACT This white paper highlights integration features implemented in EMC Avamar with EMC Data Domain deduplication storage systems
BlackBerry Enterprise Server for Microsoft Exchange Version: 5.0 Service Pack: 2 Feature and Technical Overview Published: 2010-06-16 SWDT305802-1108946-0615123042-001 Contents 1 Overview: BlackBerry Enterprise
The EMSX Platform A Modular, Scalable, Efficient, Adaptable Platform to Manage Multi-technology Networks A White Paper November 2002 Abstract: The EMSX Platform is a set of components that together provide
White Paper Scala Storage Scale-Out Clustered Storage White Paper Chapter 1 Introduction... 3 Capacity - Explosive Growth of Unstructured Data... 3 Performance - Cluster Computing... 3 Chapter 2 Current
GSX Monitor & Analyzer for IBM Collaboration Suite 1 About GSX Solutions Founded in 1996, Headquartered in Switzerland Offices in USA, UK, France, Switzerland, China 600 customers In 40 countries GSX Solutions
CrashPlan PRO Enterprise Backup People Friendly, Enterprise Tough CrashPlan PRO is a high performance, cross-platform backup solution that provides continuous protection onsite, offsite, and online for
Alliance Key Manager Solution Brief KEY MANAGEMENT Enterprise Encryption Key Management On the road to protecting sensitive data assets, data encryption remains one of the most difficult goals. A major
Case Study Cloud based Holdfast Electronic Sports Game Platform Intel and Holdfast work together to upgrade Holdfast Electronic Sports Game Platform with cloud technology Background Shanghai Holdfast Online
High-Availability Using Open Source Software Luka Perkov Iskon Internet, Zagreb, Croatia Nikola Pavković Ruđer Bošković Institute Bijenička cesta Zagreb, Croatia Juraj Petrović Faculty of Electrical Engineering
Cloud Service Model Selecting a cloud service model Different cloud service models within the enterprise Single cloud provider AWS for IaaS Azure for PaaS Force fit all solutions into the cloud service
Your consent to our cookies if you continue to use this website.