IBM and Openwave: Mx with Rich Mail on the IBM PureFlex System

Transcription

1 IBM and Openwave: Mx with Rich Mail on the IBM PureFlex System A scalable solution for a larger customer base Openwave Systems Inc. IBM Systems and Technology Group ISV Enablement April 2012 Copyright IBM Corporation, 2012

2 Table of contents Abstract...1 Introduction...1 Openwave MX stateless architecture...2 Openwave Rich Mail architecture...5 PureFlex System hardware architecture...8 Functional test environment...8 Physical component layout... 8 Load balancing... 9 Test process...10 Summary...16 Resources...17 Trademarks and special notices...18

3 Abstract This paper describes the implementation and metrics of the Openwave Mx with Rich Mail solution running on the IBM PureFlex System. The Openwave Mx with Rich Mail is a carrier grade, scalable messaging platform that provides rich messaging experience at a low cost of ownership. Openwave Mx with Rich Mail uses an industry-leading stateless architecture that scales both vertically and horizontally to support thousands of subscribers to millions of subscribers. The IBM PureFlex Platform is a tightly integrated combination of compute nodes, storage, networking, and system management all in one package. Openwave Mx with Rich Mail solution on the IBM PureFlex System platform is a tightly integrated solution that can grow to suit a large market segment. Introduction has become a critical element to business and personal communications. subscribers expect their service to be accessible anytime from a variety of devices, all managed from a single entity. In addition, subscribers require an ever-increasing demand for storage. These conditions drive service providers to implement solutions that are robust, reliable, and easily scalable to meet their subscribers needs or risk losing revenue from churn. The integration of Openwave Mx with Rich Mail on the IBM PureFlex System platform addresses the service provider challenges with a solution that is flexible, reliable, and scalable, all while being easy to administer. The benefits of the service provider include: Carrier-grade reliability to ensure there is no single point of failure in the solution Rapid introduction of new services when required, which is a solution built around virtual resources that facilitates the rapid introduction of additional resources and services to meet subscriber demands The subscriber also benefits from the integration of Openwave Mx with Rich Mail on the IBM PureFlex System platform. The benefits to the subscriber include: High-performance mailboxes: Near limitless capacity with extensive sort and search capabilities Superior web experience: Rich Mail features for fast preview, drag and drop, contextual right-click menus and in-line images. This white paper describes the architecture and implementation of this optimized, integrated solution. 1

4 Openwave MX stateless architecture and messaging in general, is an I/O-intensive application. The application puts a huge burden on the underlying storage system as a result of high amounts of I/O operations. In order to keep up with the load and still have acceptable performance, the systems in today s market are forced to use expensive storage arrays. Figure 1: Mx architecture Replacing a storage area network (SAN) array with several of the commodity disks that you normally have on your desktop can overburden the system, as the low-cost commodity storage will not be able to keep up with the I/O load exerted by the application. 2

5 Figure 2: Openwave messaging platform Openwave stateless architecture design s drastic reduction in I/O dependency, making the commodity storage disks a viable option in a commercial, production environment. The creative design and implementation of the Openwave software reduces I/O dependency and enables the use of low-cost commodity hardware without compromising reliability. This architecture provides the basis of the total cost of ownership (TCO) reduction by: Eliminating the need for expensive storage hardware (for example, SAN arrays) Eliminating any proprietary software that requires expensive licenses and support (for example, Oracle database, Symantec Veritas High-Availability software) Reducing operational expenses by making the system highly reliable and self-healing The Openwave MX stateless architecture also has a positive impact on business continuity. In order for operators to offer business continuity services in the face of disasters, the system needs to be highly reliable, with no single point of failure. The solution s geo-redundancy (GR) ensures resilience against data center outages due to natural disasters. Together, these requirements result in: Introduction of low-cost and commodity storage hardware into the messaging platform to reduce storage expense. Introduction of the Cassandra Database, a highly scalable and distributed database for storing message metadata and message body. This eliminates expensive software licenses to high availability (HA) software vendors. 3

6 Redesign of the messaging platform to make it active / active geo-redundant. This means a consumer can have simultaneous, read / write access of their message data from any georedundant site. In order to support active / active geo-redundancy (GR), the entire application stack, including the storage layer, must support geo-redundancy. Active / Active implies that a user can have simultaneous clients open against their mailbox, reading and depositing messages in two or more geographic data centers around the world, and still get a consistent messaging experience without introducing a single point of bottleneck or failure. Other challenges with active/active GR include dealing with WAN backhaul outages and variable latency on the backhaul. Active/Active GR allows customers to provide Global Server Load Balancing (GSLB)- based routing of messages without having to assign a home data center to the subscriber. In case of a natural disaster when one of the data centers suffers an outage, the GSLB can seamlessly reroute traffic to data centers at other locations with no service outages for users. Figure 3: Multi-site, geo-redundant configuration Openwave s messaging platform has many unique aspects: Built-in active/active geo-redundancy allows simultaneous access to a mailbox from multiple data centers. This enables service providers to offer business continuity services without the need for expensive HA licenses. Distributed data store allows virtually infinite incremental horizontal scalability. Commodity-based storage is used by drastically reducing I/O dependency on the storage sub- system. This allows the solution to run on low-cost commodity storage hardware, which can reduce the TCO. Quick compression to reduce storage and network bandwidth requirements. 4

7 Storage-efficient indexing for supporting the search functionality. Compared to traditional index solutions, this solution takes a fraction of the storage space for storing indexes. Redesigned Internet Message Access Protocol (IMAP) server to scale from hundreds to thousands of simultaneous IMAP users per box. Given that the industry is heading in the direction of IMAP for the future for mobile converged IP messaging, this innovative Mx redesign is on the right path. Openwave Rich Mail architecture Rich Mail is an AJAX-based web messaging interface designed to enhance the user experience by reducing fragmentation. The interface, as shown in Figure 4, consists of a familiar three-pane view with folders, an inbox, preview panes, and a widget-based view that serves as a communication dashboard. Figure 4: Rich Mail s widget-based client Rich Mail s layered architecture (Figure 5) allows its application servers to interface with the MX subsystem. 5

8 Figure 5: Rich Mail high-level architecture Rich Mail greatly improves the fragmented subscriber messaging experience, allowing subscribers to: Send, receive, and manage from the main account, and additional accounts including Gmail, Yahoo, MSN, or any other IMAP or Post Office Protocol (POP) accounts. Send, receive, and manage SMS/MMS from mobile accounts. Manage voice messages. View Facebook news feeds, share content, and add comments. View Twitter feeds and post messages. Manage contacts, calendars, and tasks. Rich Mail serves as a communications hub, allowing the user to access different social media communities concurrently, enabling effective cross-platform communications using a single interface and login. Rich Mail also offers a mobile webkit client for smart phones (Android and iphone) that provides users with a rich mobile messaging experience. Standard protocols are also supported for a native smartphone client experience: IMAP and POP for retrieval and ActiveSync for , contact, and calendar retrieval. One key element of the Rich Mail platform for the administrator is the expandability of the widget view, which is based on the Universal Widget API (UWA) specification. Widgets found in the community can be added to the catalogue and offered to users. More importantly, this allows service providers to easily add widgets that relate to their services. Rich Mail comes with the following widgets: Message: View inbox, reply to, forward, or delete any message. Contacts: Create, edit, and delete contacts. Calendar: Create, edit, or delete calendar events. News: Link to a news feed (for example: Internal news feed from the service provider about products, services, or content). Weather: Access worldwide weather information. Facebook: View and interact with your Facebook feed. Twitter: View and interact with your Twitter feed (from RM4.1). 6

9 Openwave Web 2.0 user interface Rich Mail is built on a Web 2.0 framework using AJAX constructs, providing a more dynamic experience to the user with the drag-and-drop functionality similar to a native application experience while retaining a web application s flexible deployment model and the ability to employ advertising. Startpage The startpage allows the user to quickly see what is new and interact with different online communities. Inbox viewing, as well as contact, calendar, task and InfoStore operations are performed from here. Mashup from web services are also offered (for example, Facebook), and more might be added by the administrator. The startpage can be used to extend the presentation layer to the other administrator s services such as TV content preview, voice calls, and so on. These presentation widgets are built on the UWA framework; the administrator can easily build their own to enhance the user experience. Folders To help users navigate substantial information bases, items managed by Rich Mail are stored in separate folders. There are one or more folders for each of the following items: Contact Calendar Task InfoStore document More folders can be added by the user to classify information. A folder tree display (which can be removed or resized as necessary) can be displayed on the left side of the interface to ease navigation. External accounts can be added to the folder tree by clicking on the add account button. accounts will appear as new folders in the folder tree. The module is a web application featuring a complete AJAX-based, three-pane view and native laptop client behavior. The module provides access to the core functionality in Rich Mail. An inbox view and preview pane is displayed along with the folder tree. Horizontally or vertically split views are available. The functionality is accessed from this interface for both: native and external accounts. Contact The Contact module provides contact management functionality. actions can also be accessed from the Contact module. An optional feature allows for the collection of addresses from all the messages exchanged. 7

10 Calendar The Calendar module provides complete time and event management. Monthly, weekly, and daily views are offered. Users can use typical event creation functionality, such as participant free/busy status, participant accept/decline/reject status, and more. Invitations are propagated to users outside the system through ical notifications. Standard calendar functionality, such as reminders, recurring events, and so on are also available. Task The Task module provides task management functionality, allowing users to manage the to-do list with completion state and details. InfoStore The InfoStore module provides backup for documents. Users can utilize the standard document management functions such as versioning, access control, and so on. PureFlex System hardware architecture The integrated design of the PureFlex System starts with a single rack capable of holding up to 14 nodes. Those nodes can be any combination of Intel and IBM Power nodes. In addition, built into the PureFlex System rack, are switches for Ethernet and Fibre Channel (FC) connectivity with the outside environment. The PureFlex System is designed for full redundancy, which allows for multiple switches, fans, and power supplies to assure that there is no single point of failure. Multiple PureFlex System racks can be combined to satisfy growth and larger configuration requirements. Functional test environment The functional test environment mimics the production environment in every way except scale. Each has the same number of virtual nodes, network address schema, naming conventions, and most importantly, data flow patterns. The following sections illustrate how the Openwave Mx with Rich Mail uses the flexibility of the IBM PureFlex System platform. Physical component layout The PureFlex System used for the Openwave testing used a minimal number of compute nodes. The compute nodes are using the Intel Xeon processor E at 2.70 GHz (24 total cores) with 112 GB of total memory. The Ethernet switch integrated in the chassis is a BNT 1 Gb scalable switch with 20 external ports. The compute nodes have an integrated internal connection to the switch. This switch is used to create multiple virtual Ethernet networks to isolate the TCP/IP traffic. The FC switch integrated in the chassis is a 20 port QLogic 8 Gb scalable switch with 6 external ports and 14 internal ports. The compute nodes use these integrated internal ports to connect to the switch. This switch allows for external connection to storage, such as the Storwize V7000 Unified system. 8

11 The IBM Storwize V7000 Unified system is connected directly to the PureFlex System FC switch. Load balancing A highly available architecture requires the use of load balancers as part of a complete solution. In this architecture, virtual F5 load balancers (vltm) are incorporated as part of the solution, and mimic the exact requirements of production environment to maintain high availability and load balancing across the stateless nodes. The PureFlex System chassis management module (CMM), provides management capabilities of all the elements in the chassis which include the two Intel compute nodes and the network switches. From the CMM, it is possible to perform operations, such as assigning IP addresses to the Intel compute nodes. The CMM provides a secure management network for administration of the components in the PureFlex System chassis. Figure 6 shows the components used for this Openwave test. PureFlex System rack Workload drivers Ethernet switch Intel node1 VMware ESXi server... Intel node x VMware ESXi server Fibre Channel switch 8 Gb Fibre Channel IBM Storwize V7000 PureFlex System chassis management module Figure 6: Physical component layout The workload drivers are not part of the PureFlex System chassis. These systems are not part of the Openwave solution. They are used only to generate a workload for the solution. The IBM System x 3550 servers were used as the workload drivers. 9

12 Test process The objective of this test is to demonstrate the functionality of the Openwave solution on the IBM PureFlex System platform. After the environment is installed and configured, the workload generators are used to generate a small user workload against the solution. Front-end testing: The IBM PureFlex System front-end processes (Simple Mail Transfer Protocol (SMTP), POP, and IMAP) are balanced through a load balancer. Due to the stateless nature of the protocol, the load balancer is responsible for listening for port availability on the destination server and balancing load appropriately. Middle layer testing: The IBM PureFlex System middle layer processes (Directory Master, Directory Cache, and Queue) are balanced and redundant through the software running on the IBM PureFlex System. Each component is set up in failover for its own respective process. Back-end testing: The IBM PureFlex System back-end processes - Message Store Server (MSS), storage layer) are balanced through the software running on the IBM PureFlex System. The following two fail over scenarios are tested through this paper: MSS node failure Storage layer node failure 10

13 MSS node: - Figure 7: MSS nodes The following log snippet is a successful mail deposit while all nodes are available in the IBM PureFlex System: mxin1 mta Note;MsgTrace(65/26) delivered:user=user1@ibm.com:mbox=10000:mss=mss2:from=<me@me.com>:msgid=< AAE3054.mxin1.ibm.com@[ ]>:size=435:port=5002:fromh ost= :localaddr=[ ]:sid=00ed9bf0-4f6ce :docid=47b31e4c e c1f6ce925fb During this transaction the IBM PureFlex System has hashed the user s account to mss2. 11

14 The following log snippet is a mail deposit while the mss2 node is unavailable: mxin1 mta Note;MsgTrace(65/26) delivered:user=user1@ibm.com:mbox=10000:mss=mss3:from=<you@me.com>:msgid=< AAF3054.mxin1.ibm.com@[ ]>:size=445:port=5002:from host= :localaddr=[ ]:sid=00ed9bf0-4f6ce :docid=7b36fe e ffc4ddbd33 During this transaction, the IBM PureFlex System has hashed the user s account to mss2, however, mss2 is unavailable and therefore, mss3 becomes a surrogate and accepts messages on behalf of mss2. As soon as mss2 is available again, mailboxes hashed to this mss can be processed once again mxin1 mta Note;MsgTrace(65/26) delivered:user=user1@ibm.com:mbox=10000:mss=mss2:from=<him@me.com>:msgid=< AAI3054.mxin1.ibm.com@[ ]>:size=430:port=5002:from host= :localaddr=[ ]:sid=00ed9bf0-4f6ce :docid=5dc7d e fe980a4a40f2 12

15 Storage node: The storage layer consists of storage nodes which achieve high availability by using active replication to distribute data across the storage nodes. Data on the system is eventually consistent across all the nodes. - Figure 8: Storage node layout Data is distributed (partitioned) across the nodes in the cluster using consistent hashing based and on an Order Preserving Hash function. Server and keys for the data deposit are hashed over the ring. In Figure 8the figure 9 below, the ring consists of three storage nodes with a replication factor (RF) of 3. If one of the three nodes becomes unavailable, the solution continues to operate as intended. In the example of a write to the storage nodes, the data stored in a primary node (Node A) is hashed (h(key 1)), and then replicated two more times (Node B, Node C) to have a total of three copies in the storage layer. 13

16 In the example of a read to the storage nodes, the data is hashed (h(key 2)), and read from the appropriate home node (Node C). h(key 1) A RF = 3 C B h(key 2) Figure 9: Hash ring and data distribution 14

17 The following log snippet shows an active POP session accessing a mailbox while one of the storage nodes is unavailable. Escape character is '^]'. +OK InterMail POP3 server ready. user user1 +OK please send PASS command pass user1 +OK user1 is welcome here list +OK 1 messages retr 1 +OK 434 octets Return-Path: <me@me.com> Received: from [ ] by mxin1.ibm.com (InterMail vm ) with SMTP id < AAM26243.mxin1.ibm.com@[ ]> for <user1@ibm.com>; Tue, 27 Mar :36: subject: professor is late Date: Tue, 27 Mar :36: Message-Id: < AAM26243.mxin1.ibm.com@[ ]> that's all there is!. Connection closed by foreign host. 15

18 Summary The PureFlex System platform is designed for flexibility with the simplicity of an appliance with integrated expertise tuned to for a task. In a single package, heterogeneous hardware is integrated with networking and storage. It is highly reliable with no single point of failure. Openwave offers a carrier-grade messaging and social platform that has been deployed at over 50 broadband and mobile operators across the globe. Service provider customers are facing unprecedented storage demand from their consumers, as infinite, searchable, always available mailboxes are what consumers are expecting. This consumer demand is driven by ever increasing use of mobile device, ubiquitous web access, and increase in average message size. All these result in deep mailboxes with extreme storage needs. The demand for business continuity services that protect subscriber data and make it available under all circumstances requires service providers to make additional investments in storage. The cost of storage with current architecture prevailing in the marketplace is prohibitive for the operators. Service providers need commodity storage based, low-tco solution that can offer carrier-grade reliability with desired business continuity, even in face of data center wide disasters. Openwave s newly designed messaging platform does exactly that by introducing commodity storage based, active/active geo-redundant solution. The Openwave Mx with Rich Mail has been fully tested by IBM and Openwave on the IBM PureFlex System platform, and has demonstrated the functionality of this solution. 16

19 Resources The following websites provide useful references to supplement the information contained in this paper: IBM Systems on PartnerWorld ibm.com/partnerworld/systems IBM Expert Integrated Systems ibm.com/systems IBM Redbooks ibm.com/redbooks Openwave Mx with Rich Mail on IBM PureFlex System 17

20 Trademarks and special notices Copyright IBM Corporation References in this document to IBM products or services do not imply that IBM intends to make them available in every country. IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol ( or ), these symbols indicate U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at Java and all Java-based trademarks and logos are trademarks or registered trademarks of Oracle and/or its affiliates. Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. Intel, Intel Inside (logos), MMX, and Pentium are trademarks of Intel Corporation in the United States, other countries, or both. UNIX is a registered trademark of The Open Group in the United States and other countries. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Other company, product, or service names may be trademarks or service marks of others. Information is provided "AS IS" without warranty of any kind. All customer examples described are presented as illustrations of how those customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics may vary by customer. Information concerning non-ibm products was obtained from a supplier of these products, published announcement material, or other publicly available sources and does not constitute an endorsement of such products by IBM. Sources for non-ibm list prices and performance numbers are taken from publicly available information, including vendor announcements and vendor worldwide homepages. IBM has not tested these products and cannot confirm the accuracy of performance, capability, or any other claims related to non-ibm products. Questions on the capability of non-ibm products should be addressed to the supplier of those products. All statements regarding IBM future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only. Contact your local IBM office or IBM authorized reseller for the full text of the specific Statement of Direction. Some information addresses anticipated future capabilities. Such information is not intended as a definitive statement of a commitment to specific levels of performance, function or delivery schedules with respect to any future products. Such commitments are only made in IBM product announcements. The information is 18

21 presented here to communicate IBM's current investment and development activities as a good faith effort to help with our customers' future planning. Performance is based on measurements and projections using standard IBM benchmarks in a controlled environment. The actual throughput or performance that any user will experience will vary depending upon considerations such as the amount of multiprogramming in the user's job stream, the I/O configuration, the storage configuration, and the workload processed. Therefore, no assurance can be given that an individual user will achieve throughput or performance improvements equivalent to the ratios stated here. Photographs shown are of engineering prototypes. Changes may be incorporated in production models. Any references in this information to non-ibm websites are provided for convenience only and do not in any manner serve as an endorsement of those websites. The materials at those websites are not part of the materials for this IBM product and use of those websites is at your own risk. 19