SmartSafe Database Replication as a Service (DbRaaS), is a managed service offering provided by Smart Associates which utilises the database replication features of our Smart Management Frameworks (SMF) product. It is designed to improve the performance, availability and recoverability of one or more of IBM®’s Pure Data™ for Analytics (aka Netezza) or Netezza Performance Server for IBM Cloud Pak for Data appliances.
SmartSafe allows customers to replicate their Netezza databases in near real-time (you choose how frequently you want to replicate changes between systems, according to workload, data volatility, bandwidth, and Recovery Point Objectives (RPO) considerations). Mean Time to Recovery (MTTR) can be measured in the minutes it will take to repoint your DNS servers to the DR system; but even without a disaster, database replication potentially enables customers to divide their data, users, and workloads between multiple appliances for improved performance, concurrency and throughput.
If you're upgrading your existing Netezza appliances to IBM NPS for Cloud Pak for Data, the migration process is substantial and not without risk. The recommended approach often requires engaging costly professional services to undertake this task, which you can avoid by using SmartSafe.
SmartSafe allows you to undertake an effortless and virtually risk-free migration to your new platform as part of the service. In a nutshell, you run your old and new Netezza systems in parallel, replicating the databases.
So no drawn out migration plan with rehearsals and outages - when you are satisfied that the new solution is fully functional, you cut over. You can even migrate your users gradually to the new system.
The IBM recommended approach to migrating to CP4D involves using the nz_migrate command. Even using this approach, planning and executing your migration is a complex task and there are a number of variables that will determine the effort and duration of the migration. We have developed an estimating tool that will show you what you can expect based on your particular system.
Businesses can save more than $1million per downtime event (and often much more) by reducing Mean Time to Recovery (MTTR) to minutes to using SmartSafe's database recovery functions.
While the cost of downtime during data recovery will vary for each customer, and each customer should have a good idea of what their own cost is, Gartner have calculated that the average cost is over $300,000 per hour. Other similar methods give a range starting from at least $140,000 per hour even for smaller firms. In our experience getting a 20Tb+ Netezza system back on-line and up to date can take a minimum of eight hours when restoring from a full backup (or significantly longer depending on the number of streams that are used, or if multiple differential and cumulative backups also need to be restored). SmartSafe will reduce this to minutes, reducing expensive downtime to a fraction, assuring business continuity and enabling your business user and analytics workloads to continue with the minimum of disruption.
Netezza Replication Services is a feature of Netezza that allows Netezza databases to be replicated between a master and one or more secondary nodes. SmartSafe, on the other hand, is a managed service provided by Smart Associates that allows multi-master Netezza database replication with a number of enhanced features.
Here is a comparison between the capabilities of each:
|Header||SmartSafe||Netezza Replication Services|
|Appliance Cross-Type (CP4D/NPS) Support||Yes||No|
|Multi Master Support||Yes||No|
|Multi Secondary Support||Yes||Yes|
|Same Database in Different Replication Sets||Yes||No|
|Partial Database Restore (Selected Tables)||Yes||No|
|Partial Table Data Restore (via customisable data filter)||Yes||No|
|Cross Database View Fixing||Yes||No|
|Out-of-the-Box Metadata Replication Setup||Yes||Yes|
|Own Replication Metadata Database||Yes||Yes|
|Data Auto-Healing Mechanisms||Yes||No|
|Built-In Integrity Checks||Yes||Yes|
|Replicating Databases With Different Name||Yes||No|
|Configurable Data Resynchronisation Frequency||Yes||No|
Although related, they are two different services. Disaster recovery as a service (DRaaS), which is currently not offered by Smart Associates, is the hosting, or replication, of physical or virtual servers by a third party to provide failover in the event of a natural catastrophe, power failure or any other type of disaster that disrupts the business. Organisations may opt for a full disaster recovery solution, or one that restores only some business critical systems. Whether or not the Data Warehouse is included in the disaster recovery solution, SmartSafe Database Replication as a Service is useful not only in the event of disaster, and businesses can derive value from the service continuously.
Apart from the peace of mind and reassurance that comes with knowing that there is an up-to-date copy of all data stored in a different location, businesses can divide their workload between their primary and recovery data centres to get improved performance, concurrency and throughput for all their users' database queries. Furthermore, SmartSafe is multi master, meaning that you don't have to have one primary system and one backup system, you can have multiple primary systems. This allows an enterprise to divide up its source systems, applications and users between two or more Netezza systems and keep all of them synchronised.
When an enterprise maintains a replicated Netezza architecture, the replication that occurs between the primary and secondary systems is typically uni-directional, i.e. the primary system is fed by the source systems, and the secondary systems are mirror copies of the primary system. If the primary system fails, a secondary system assumes the role of primary, DNS changes are made to redirect the feeds and the user connections, and business carries on as usual until the other system is fixed and roles are reversed again.
Traditional Netezza Database Replication
SmartSafe database replication is bi-directional (BDR) and allows all Netezza nodes to be primary systems.
An example of this is where you have two or more systems and each of them can be the primary system for a set of users, one or more applications and one or more source systems, and you can replicate the data between them all. This gives you a natural way of dividing your users across your infrastructure whilst still being able to have replication across all those different systems. Moreover, in a multi-master scenario, each master also potentially acts as a Disaster Recovery site for one or more of the other masters.
Multi-master bi-directional replication (BDR) as provided by Smart Management Frameworks
If you are thinking of buying a new Netezza Performance Server for IBM Cloud Pak for Data system or of migrating to Netezza in the cloud, Smart Management Frameworks SmartSafe can be used to automate the migration between the old and the new environments.
Replicating the existing Netezza databases to the new environment allows the migration to be performed incrementally and in a highly controlled manner. It allows for a dual run of both old and new systems side by side which means that you can control when the cutover occurs between the old and the new.
For customers wanting to try out the cloud service, SmartSafe will allow you to do a dual-run whilst doing an evaluation as part of the migration process. So if you are not sure if the Cloud solution will cope with the volume of data, queries and users and be better and more reliable than the on-premise option, it could potentially be used to replicate some or all of your existing data to that cloud option. It provides a safe way of testing the new netezza platform without having to take a leap in faith, keeping existing systems in place until you are ready to do the transition and the switch over, without it having to be a big-bang approach.
Many customers maintain a separate Netezza appliance as a development environment, and this system will typically have much less capacity than the production system. Maintaining a subset of data on dev can be quite a fiddly task, especially if you want the data accurately reflect the production environment. With SmartSafe you can replicate a subset of your production data to your dev environment so that it contains real world data that is accurate and up-to-date and small enough to fit on the dev environment. In exactly the same way you could create a separate test environment, and replicate a sub-set of production data to it.
So if, for example, you are constrained by disk space and so only want the last six months of transactional data on dev and test, SmartSafe can maintain that rolling window of six months of historical data for you. Alternatively, it can replicate a percentage of records using a percentage algorithm to determine the data that gets replicated. This is a much better option having to using a WHERE constraint or applying filter to the source data
SmartSafe works with all combinations of versions of Netezza systems - from the PureData for Analytics Nx00x range, to IBM’s latest generation Netezza on System or Netezza on Cloud offerings.
By choosing SmartSafe you will be able to replicate your old and new Netezza systems and avoid the problems that are invariably associated with upgrading to new technologies. Why is this? Because you can run your old and your new systems in parallel and choose the moment you cut-over with a few simple steps, obviating the need for extended system downtime and drawn-out post-migration user acceptance testing.
Mean Time to Recovery is quite simply, “how long will it take to recover our system to being fully operational after a failure”. It is sometimes also referred to as Recovery Time Objective.
MTTR is important to the business as it is regarded as a measure of how fit for purpose your Netezza disaster recovery system is. It implies that a full disaster recovery test has been conducted and that this time to recovery has been achieved, rather than just being an objective. If the MTTR has not been tested, your business runs the risk of losing their Netezza system for a considerable period of time which may have a serious impact on business continuity.
Recovery Point Objective determines how much data the business can afford to lose after an outage. So, does the data have to be right up to date, or can the business tolerate losing some of it? It factors in that the database transactions that have occurred since the most recent backup will also need to be restored if the RPO is quite a short one, and this may impact the Mean Time to Recovery.
For customers wanting to migrate from an alternative platform to Netezza, or vice versa, we have a solution called Smart Data Frameworks which will not only automate the migration but can also keep trickle-feeding changes from the Netezza database to the new target database until you are ready to switch over.
So, in the same way that DR as Service can be used to do a parallel run of an old and new Netezza platform, SDF will allow you to do a parallel run between an old Netezza and a new non-Netezza system, as long as you have primary keys declared in your tables. Accordingly, this is a form of replication but it is not designed to be permanent, but rather to assist with the migration process.