SAP Technology SAP Adaptive Server Enterprise, Enterprise Edition, High-Availability and Disaster Recovery Options, to Boost Overall Performance Table of Contents 2 Introduction 2 A One-Two Punch: Planned and Unplanned Downtime 7 SAP Replication Server 8 Warm Standby Replication with SAP ASE 3 Legacy Approaches to High- Availability Disaster Recovery 5 SAP Adaptive Server (SAP ASE) Enterprise: An Integrated Approach to High-Availability Disaster Recovery 6 Powerful High-Availability Disaster Recovery Solutions 9 SAP ASE, Enterprise Edition, High-Availability and Disaster Recovery Options 10 SAP ASE Cluster Edition 13 Achieving Continuous Availability with SAP ASE
In today s fiercely competitive business environment, an increasing number of companies expect their IT departments to provide anytime, anywhere access to mission-critical services. Yet these same companies also charge IT to do more with less, demanding the lowest total cost of ownership (TCO) possible. INTRODUCTION Unfortunately, many decision makers do not fully understand the trade-offs that exist between higher availability and cost containment. For instance, while 99% availability translates into a surprising 1.68 hours of downtime per week, that downtime drops to just over 10 minutes per week at 99.9% availability. This seemingly small uptick in availability comes with a marked drop in downtime, but it also comes at a price. There are numerous challenges to delivering the service availability levels that enterprises demand today. But the consequences of not addressing these challenges effectively can be significant, including reduced corporate agility, unmet service-level agreements, unacceptably high operational costs, user dissatisfaction, and even lost business. Many organizations have successfully addressed these challenges with SAP Adaptive Server Enterprise (SAP ASE) a high-performance transactional database that supports missioncritical business applications. Its range of robust options enables IT teams to deliver the levels of availability, disaster recovery, resiliency, and economy their enterprises require to meet business needs both today and in the future. A ONE-TWO PUNCH: PLANNED AND UNPLANNED DOWNTIME Of the two main types of system downtime, planned downtime represents more than 90% of all outages. In the past, this was a manageable issue for IT, since most systems were never expected to be online continuously, and scheduled maintenance windows were a standard business practice. But today, planned downtime including hardware, software, database maintenance, application upgrades and updates, and environment migration is a major impediment to meeting the always-on availability demanded by enterprise users. Despite the fact that unplanned downtime occurs much less frequently than planned downtime, the ramifications of these unexpected outages are frequently more disruptive to enterprise operations since they can happen during peak times. Common causes of unplanned downtime include: Database host or software failures Probably the best-understood and most frequently mitigated risk, thanks to clustering, hot and warm failover, and other technologies Sitewide failures Typically caused by physical events at a site, such as electrical failure, fire, or flood, but can also result from broad network failures 2 / 13
Data access failure Caused by unintended or willful data corruption, or by failure of primary data stores; two notable instances: a catastrophic storage failure in 2013 during planned maintenance downtime that delayed unemployment checks for the entire state of Oregon, and a major credit card issuer that suffered a failure that halted normal business operations at the company for three days Denial of service (DoS) The most pervasive cause of unplanned downtime, DoS can be the result of malicious attacks, but it can also be caused by a code defect in, or sudden surge in demand for, a legitimate application; an example: a simple distributed DoS attack recently caused Sony Corporation s PlayStation Network and Sony Entertainment Network to be unavailable to millions of users Whatever the cause, an outage of sufficient duration or severity not only disrupts operational schedules, it can also permanently damage the reputations of affected service providers. 90% Of all systems outages are planned downtime a major impediment to meeting the always-on availability demanded by enterprise users. LEGACY APPROACHES TO HIGH-AVAILABILITY DISASTER RECOVERY Enterprise IT decision makers have long invested in defenses against system threats, but many of these protections come with significant tradeoffs in cost, effectiveness, and applicability to specific environments. Some of the most common starting points for achieving higher availability and resiliency levels for enterprises include: Hardware clustering Also known as operating system (OS) clustering, this technique connects two or more physical servers and controls them with a common OS. The OS runs on dedicated hardware and monitors the cluster, determines when failover from a failed system (or node) is needed, and assigns the failed node s work to a functioning node. Such clusters can be active-active, with all nodes running applications and available for failover duty, or active-passive, with some nodes kept idle until they are needed for failover. Hardware clustering typically requires specialized hardware and modification of applications to make them cluster-aware. Application clustering With this technique, clustering software is installed on each server, and the servers collectively handle loadbalancing and failover-related tasks. This approach makes adding and removing servers relatively easy, and requires no specialized hardware. As a result, application and software clustering are easily scalable and generally less expensive and more popular than hardware and OS clustering. 3 / 13
Disk block replication Largely popularized by storage area network (SAN) vendors, this approach has also been implemented by storage software providers and involves synchronous or asynchronous duplication of blocks of data across disks in two or more sites. Disk block replication is often used in conjunction with a redundant array of independent disks (RAID) deployment. Virtualization This technology creates independent virtual computing and storage resources from single or clustered physical versions of those resources. Virtualization can be accomplished in multiple ways. While not an availability solution per se, virtualization can ease and speed restoration of failed services in clusters and limit the effects of failed applications on other applications that are still running. Transaction log shipping This methodology automatically and regularly sends a copy of the transaction log files affiliated with a database from a primary server to a backup server. Although easy and inexpensive to implement, failover to the backup server is not automatic. In addition, the batch-shipping nature of this solution poses a risk of data loss. While each of these techniques can help improve service availability, disaster recovery, and resiliency, none alone is sufficient to address today s rigorous demands for continuous availability. When evaluating alternative approaches for continuous availability, most features and functions can be grouped into four categories: Fault tolerance the solution s ability to prevent or protect against outages Application impact the solution s effect on application availability, functionality, performance, and usability Implementation considerations the solution s effect on incumbent resources, skills, and staffing Recovery effort how rapidly and automatically the solution can recover from a fault, failure, or outage In addition, TCO is a major factor when evaluating high-availability disaster recovery (HADR) solutions. This is especially true for approaches that require the purchase of new hardware or the repurposing of legacy hardware. Such approaches almost always result in higher operational expenses and greater infrastructure complexities, two obvious roadblocks to low TCO and high availability. TCO is a major factor when evaluating high-availability disaster recovery (HADR) solutions. This is especially true for approaches that require the purchase of new hardware. 4 / 13
SAP ADAPTIVE SERVER ENTERPRISE: AN INTEGRATED APPROACH TO HIGH-AVAILABILITY DISASTER RECOVERY Initially designed to meet the exacting standards of both Wall Street and the U.S. military, SAP ASE brings high availability to mission-critical, highvolume, high-throughput applications worldwide. Today, enhancements to the software continue to be made with the following goals in mind: Prevent system outages caused by maintenance activities or server reconfiguration Reduce system outages due to DoS as a result of runaway queries, DoS network attacks, or system overloads from other causes Provide rapid recovery when a failure does occur and ensure that applications deemed most critical by the business are restored first Isolate data corruption to the affected data blocks to avoid having to take entire databases or critical tables offline and to minimize data loss and operational disruption, even while corrupted indexes are being rebuilt 2.5 terabytes Per hour backup speeds have been measured on databases larger than 10 TB on a single file system. To achieve these aggressive goals, SAP ASE has been equipped with many advanced features, including: Enhanced backup and restore SAP ASE provides a high-speed online backup for no additional licensing costs. No extra configuration effort is required beyond designating the desired compression and parallelism characteristics for each backup. Backup speeds have been measured at more than 2.5 terabytes (TB) per hour on databases larger than 10 TB on a single file system. SAP ASE includes point-in-time recovery functionality, a backup application programming interface (API) for custom integration, backup compression, backup access and password control, and optional support for integration with third-party backup solutions. SAP ASE can also be restored online. Configurable fault isolation and recovery Database administrators can isolate a data corruption repair down to the page level. This means that they can repair corrupted data while the rest of the system remains online and available. SAP ASE provides for prioritized recovery and supports the use of a standby system for reporting purposes. These features help businesses get back to full operation more quickly and diagnose, resolve, and prevent faults more effectively. Resource virtualization SAP ASE was built on a virtual server architecture model. The solution takes a multifaceted approach to process and resource isolation, allowing IT decision makers to match availability levels to resource criticality more closely and use resources more efficiently. 5 / 13
Online maintenance Nearly all of the configuration parameters of SAP ASE (including memory allocation) can be changed while the system is online, minimizing operational disruption during maintenance. In addition, SAP ASE includes an integrated job scheduler and templates for frequently required maintenance tasks. These features allow needs-based maintenance to be scheduled automatically without the additional cost of separate job scheduling software. Full database encryption Comprehensive data encryption enhances security, which contributes directly to higher availability. SAP ASE supports full database encryption, even while users are accessing data. All of these features contribute to uptime performance that significantly exceeds that of alternative HADR solutions. In fact, market analyst IDC found that downtime was reduced by 89% for organizations running SAP ASE versus competitive databases. 1 POWERFUL HIGH-AVAILABILITY DISASTER RECOVERY SOLUTIONS Looking beyond the robust availability features delivered in SAP ASE, SAP offers a number of powerful solutions for heightened levels of availability and disaster recovery (see Figure 1). 1. Carl W. Olofson and Randy Perry, Reducing Cost While Simplifying Administration: Monetizing the Benefits of SAP ASE, IDC, Sponsored by SAP, June 2014. Figure 1: High-Availability Pyramid of SAP Adaptive Server Enterprise (SAP ASE) SAP ASE Cluster Edition SAP ASE data replication SAP ASE inherent features Hardware availability solutions Hardware redundancy (for example, RAID, network cards) Hardware resiliency features As one goes up the pyramid in Figure 1, risk mitigation increases and outage duration and recovery times decrease. With traditional tools, solutions higher in the stack depend on lowerlevel solutions. For example, OS clustering for enhanced availability is often contingent upon the redundancy of hardware, such as network cards or disk drives. High-availability solutions based on SAP ASE are more independent. A case in point: clusters running SAP ASE do not rely on data replication. But by combining clustering based on SAP ASE with data replication, greater protection from a wider range of risks becomes possible. For years, SAP ASE has supported multiple HADR solutions that enable IT decision makers to fine-tune their deployments in ways that maximize availability and business benefits while minimizing TCO. 6 / 13
SAP REPLICATION SERVER As noted above, data replication plays an important role in several HADR solutions based on SAP software including warm standby replication functionality of SAP ASE and the SAP ASE, enterprise edition, high-availability option, both of which are described below. SAP Replication Server, a leading replication solution, has been available from SAP for the past 17 years, with more than 30,000 deployments worldwide. Designed to meet the exacting demands of Wall Street and multinational 24x7 enterprises, SAP Replication Server runs a software process alongside the database server. Unlike solutions that require the database to be programmed to alert the software to changes, SAP Replication Server works nonintrusively by automatically monitoring and mining the database s transaction log for critical transactions. SAP Replication Server provides enterprises with: Continuous backup An ongoing, low-latency mechanism for moving data from source to target system helps ensure that backup systems are always up-to-date. Transactional consistency Unlike systems that copy disk blocks or live database pages, SAP Replication Server uses the actual database transaction as the base unit, so the transaction data at the target site is consistent with its source. Nonintrusive operation Changes are examined at the database log level rather than by polling the source database, allowing primary data servers to continue to function at peak performance. Easy integration Setup and configuration are performed without requiring changes to the application code, which simplifies implementation. Heterogeneous data source support SAP Replication Server supports data compatibility among deployments of SAP ASE, Oracle Enterprise, Microsoft SQL, and IBM DB2. Long-distance backups Synchronized copies of data can be created over wide-area networks (WANs) hundreds or thousands of miles away. Corruption protection Data servers, disk drives, and other computer components sometimes experience glitches that can leave incorrect data on disk. Since SAP Replication Server copies only actual transactions, it does not accidentally copy corrupt data from the source to the backup server. HIGH-AVAILABILITY DISASTER RECOVERY FOR SAP BUSINESS SUITE SAP Replication Server provides a near-zero downtime disaster recovery solution certified for SAP Business Suite software running on SAP Adaptive Server Enterprise (SAP ASE). SAP Replication Server aligns with the road map for SAP Business Suite and is included with the runtime license for SAP ASE for SAP Business Suite. 7 / 13
WARM STANDBY REPLICATION WITH SAP ASE Warm standby replication based on SAP ASE is probably the high-availability solution most familiar to current customers of SAP ASE. By using SAP Replication Server, it provides support for organizations with great distances between sites and mitigates many of the risks related to planned outages a level of support that hardware solutions can t match. This proven solution supports master database replication keeping logins, roles, and systemwide security in sync as well as heterogeneous databases and encrypted data. Standby sites can be fully leveraged for running reports, protecting the primary system from overloads during peak reporting periods. The warm standby replication functionality of SAP ASE provides comprehensive details about outages and recoveries, which further improves system management, availability, and TCO. Bidirectional data flow ensures that IT can synchronize audit trail data from reports run on standby servers with the primary active system. The warm standby replication functionality can be deployed at a single site or at multiple standby sites. An example of a multiple standby site deployment is shown in Figure 2. Figure 2: Warm Standby Replication at Multiple Standby Sites Active/Primary Site Standby Site 1 Primary SAP Adaptive Server Enterprise (SAP ASE) Standby SAP ASE Software binaries Software binaries System databases System databases Active database SAP Replication Server SAP Replication Server Standby database Standby Site 2 Standby SAP ASE Software binaries System databases SAP Replication Server Standby database 8 / 13
In such a deployment, one standby system can be used to implement application upgrades that require schema changes, hardware platform migrations, major database upgrades, or recovery over long distances. Meanwhile, the second standby system is kept in sync with the primary active system for total data protection. In addition, in this configuration, data can flow bidirectionally among primary and standby servers full time. This means that IT teams can use application provisioning to host multiple applications on different servers or to migrate applications in phases, maximizing availability while minimizing downtime. Optionally, the SAP Open Switch application can be employed to automate fault detection and failover between primary and standby sites, further shortening periods of downtime. SAP ASE, ENTERPRISE EDITION, HIGH-AVAIL- ABILITY AND DISASTER RECOVERY OPTIONS The high-availability option and disaster recovery option of SAP ASE, enterprise edition, were specifically developed to provide for zero data loss or near-zero data loss, as well as other distinct advantages over competitive offerings. These advantages are due in part to the hybrid data replication approach used, as shown in Figure 3. Data replication is typically performed in three phases. Data is captured from a primary system, then distributed and replicated to the standby system. SAP Replication Server software supports the high-availability and disaster recovery options of SAP ASE, enterprise edition, which capture and forward transactions through synchronous, near-synchronous, and asynchronous protocols. Transport and application of changes to the standby system employ the traditional asynchronous approach of warm standby replication. Figure 3: Data Replication with SAP ASE HADR Mode Figure 3: Data Replication with SAP Adaptive Server Enterprise, Enterprise Edition Capture Distribute Apply Primary Standby Synchronous, near-synchronous, or asynchronous Asynchronous 9 / 13
The high-availability and disaster recovery options of SAP ASE, enterprise edition, offer compelling IT and business benefits, allowing enterprises to: Greatly extend the distance by which primary and recovery sites can be separated, increasing deployment flexibility and data protection Enable zero data loss if the primary SAP Replication Server is located elsewhere, even after a catastrophic failure of the primary physical site Eliminate data loss, even if a standby server is offline when a primary server fails, by storing information about transactions and changes, and then sending that information when the standby server resumes operation Support different schemas for primary and recovery sites, easing accommodation of planned downtime Keep multiple standby systems in sync without jeopardizing performance or transaction rates at the primary system. SAP ASE CLUSTER EDITION Another important HADR technology is SAP ASE Cluster Edition, which provides continuous availability levels that approach 100%. This mission-critical edition supports a shared-disk clustering architecture in which multiple servers running SAP ASE cooperate to manage the same data and provide concurrent access to one or more databases from multiple server hosts. With SAP ASE Cluster Edition, clustered database systems present themselves as a single system to applications. If one or more nodes in the cluster should fail, other nodes in the cluster will carry on without interrupting business. Unlike traditional active-active clustering solutions, SAP ASE Cluster Edition can handle multiple node failures with faster failover and failback. In addition, the edition s underlying interconnect layer eliminates the overhead typically associated with clustered database solutions. These features enable IT to offer higher levels of service availability without incremental effort or risk. SAP ASE Cluster Edition is ideally suited to environments with any of the following characteristics: Very low tolerance for downtime Business requirements calling for nearly instantaneous failover of database processing upon an unplanned outage Small or nonexistent maintenance windows for planned downtime activities like upgrading hardware or patching operating systems Requirements for a more cost-effective or application-aware, high-availability solution than is possible with traditional cluster architectures Support for a large number of small applications, many of which currently run on dedicated servers 10 / 13
SAP ASE Cluster Edition brings multiple strengths and benefits to such environments: Protects application service levels SAP ASE Cluster Edition brings system uptime to near-100% levels by handling multiple node failures with fast recovery times. It enables IT to put standby servers to work handling database requests during peak loads. Minimizes the effect of maintenance activities on operational performance System administrators can migrate database connections to standby servers to take other nodes offline for routine maintenance. This effectively enables administrators to split database maintenance windows into smaller segments and to minimize their impact on system performance. Simplifies implementation of virtual workloads SAP ASE Cluster Edition provides virtualized resource management software. This maximizes availability by enabling applications to failover and failback transparently during both planned and unplanned outages. It improves the cluster edition s ability to cope with higher-thannormal peak load requirements by enabling online migration of workloads from one node to another. Configuration is simple, flexible, powerful, and driven by sophisticated business rules. Provides a load-balancing system automatically and transparently Unlike architectures that require applications to be explicitly partitioned across different nodes, SAP ASE Cluster Edition automatically distributes workloads across multiple nodes transparently, based on specific business rules. It enables automatic activation of standby servers within the cluster as needed to process excess database requests during peak load times. One common deployment scenario for the cluster edition of SAP ASE is a multinode, activepassive configuration. This enables a number of separate database servers to share a passive standby node. This model greatly improves the economies of a cluster by amortizing the cost of the additional standby across a larger set of systems running on the cluster. While economical, this active-passive approach poses a risk of overloading individual standby nodes if multiple nodes fail simultaneously. 11 / 13
An alternative multinode, active-active consolidation configuration often delivers the greatest business value. In this configuration, SAP ASE Cluster Edition provides a shared platform upon which multiple applications and their associated databases can be consolidated. An example of such a configuration is shown in Figure 4. Figure 4 shows how eight servers running SAP ASE, each on its own host, can be consolidated onto half as many hosts while leaving sufficient capacity on each host to provide failover support for other instances. The configuration supports transparent failover of application connections to the database, and faster database failover than possible with traditional high-availability alternatives. These features mean that IT can provide high-availability support to applications and their users without additional cost or effort. Figure 4: SAP Adaptive Server Enterprise Cluster Edition (SAP ASE Cluster Edition) BEFORE: Eight servers of SAP ASE running on underutilized platforms A B C D E F G H A B C D E F G H AFTER: SAP ASE Cluster Edition SAP ASE Cluster Edition with four instances A C E G B D F H Standby Standby Standby Standby Storage area network A B C D E F G H 12 / 13
Figure 5: SAP Adaptive Server Enterprise Cluster Edition (SAP ASE Cluster Edition) Moving application B from one instance to another SAP ASE Cluster Edition with four instances A B Standby C E G D F H Standby Standby Standby Storage area network A B C D E F G H In a scenario like the one shown in Figure 5, SAP ASE Cluster Edition advanced management features come into play. Application A experiences a higher-than-average peak load. Rather than impact the service-level agreements to both this application and other applications on that instance (B), a database administrator can choose to move the running application B to another host temporarily. This provides extra capacity to application A while ensuring application B isn t disrupted. By allowing database administrators to associate specific cluster nodes with applications, IT can allocate capacity based on business priorities and dynamically shift resources among applications as needed. ACHIEVING CONTINUOUS AVAILABILITY WITH SAP ASE The need for high even continuous availability has never been greater, due to the always-on nature of business today. Overcoming the challenges of planned and unplanned downtime requires robust and integrated HADR solutions that protect the business while delivering sustainable TCO. SAP Adaptive Server Enterprise helps companies meet today s demand for continuous availability for their mission-critical operations. Unlike many traditional solutions, SAP ASE offers superior inherent resiliency paired with proven, flexible technologies that address high-availability and disaster recovery needs. Contact your SAP representative today to learn more about SAP ASE and how it can help your IT department deliver maximum availability and performance for your company s most critical resources. 13 / 13 37412enUS (15/5)
No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of SAP SE or an SAP affiliate company. SAP and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of SAP SE (or an SAP affiliate company) in Germany and other countries. Please see http://www.sap.com/corporate-en/legal/copyright/index.epx#trademark for additional trademark information and notices. Some software products marketed by SAP SE and its distributors contain proprietary software components of other software vendors. National product specifications may vary. These materials are provided by SAP SE or an SAP affiliate company for informational purposes only, without representation or warranty of any kind, and SAP SE or its affiliated companies shall not be liable for errors or omissions with respect to the materials. The only warranties for SAP SE or SAP affiliate company products and services are those that are set forth in the express warranty statements accompanying such products and services, if any. Nothing herein should be construed as constituting an additional warranty. In particular, SAP SE or its affiliated companies have no obligation to pursue any course of business outlined in this document or any related presentation, or to develop or release any functionality mentioned therein. This document, or any related presentation, and SAP SE s or its affiliated companies strategy and possible future developments, products, and/or platform directions and functionality are all subject to change and may be changed by SAP SE or its affiliated companies at any time for any reason without notice. The information in this document is not a commitment, promise, or legal obligation to deliver any material, code, or functionality. All forward-looking statements are subject to various risks and uncertainties that could cause actual results to differ materially from expectations. Readers are cautioned not to place undue reliance on these forward-looking statements, which speak only as of their dates, and they should not be relied upon in making purchasing decisions.