The EDW evolution continues and it is Bigger, Faster and Better !
As some of you may have heard, during the week of November 6th, 2012 HP and Microsoft has announced the next generation of the Enterprise Data Warehouse (EDW) Appliance.
Let’s take a moment to discuss some major changes and feature enhancements that increases the EDW appeal and makes it more appealing, with greater flexibility and scalability than the EDW V1 appliance.
Without resorting to marketing hype, the EDW appliance is a great, cost effective and easy to manage way to complement any organizations structured big data requirements.
Data may be loaded by using traditional ETL tools or by using Hadoop connectors to provide a comprehensive BI solution for executives, power users or more traditional end users.
The use of a hub and spoke architecture allows power users to access state of the art data mining / advanced analytic tools (such as Vertica or SAS, etc.) without interfering with executives or a large number of regular users who are more interested in running executive dashboards, EXCEL, PowerPivot, OLAP queries or reporting operations. This type of architecture insures excellent and consistent performance for all BI users.
A great example of flexibility relates to the fact that EDW V2 will allow customers to use their own, existing or new hardware, to perform backup operations. EDW V1 required customers to purchase a backup node and its respective storage. While EDW V2 still requires a backup system we simply specify backup minimum requirements and best practices.
HP, also, plans on offering an “HP Appsystem” for EDW backup, but there is no requirement to purchase it as part of the EDW V2 appliance.
In a similar vein, the Landing Zone function is still a requirement for an EDW appliance. But, the customer may use his own existing ETL tier or purchase an optional “HP Appsystem” to support Landing Zone / ETL functionality.
EDW V2 vs. EDW V1 technical differences and scalability
In an effort to provide greater flexibility, scalability and availability Microsoft enhanced the SQL Server Parallel Data Warehouse (PDW) software by incorporating Hyper-V, storage provisioning, mirroring, and Windows Server 2012 storage Spaces. Coupling these enhancements with HP’s latest servers, more cost effective “direct attach” SAS JBOD storage and FDR InfiniBand allows the EDW to be more highly available while scaling up to 64 nodes.
In addition to increased CPU processing power (16 cores per EDW V2 compute node vs. 12 cores for EDW V1 compute node), the EDW V2 contains 256GB of memory per server vs.96GB for EDW V1. The EDW V2 appliance also contains 35 disks per EDW V2 Compute node vs. 11 or 24 disk options on the EDW V1 appliance. This will allow customers to grow the EDW V2 to support up to 5 PB (Petabytes) of data.
It is important to note that HP and Microsoft try to insure a balanced CPU to Disk ratio in an effort to use all the configured hardware more effectively without inherent bottlenecks. Hence, since EDW V2 has more, faster cores we increased the number of disks per compute node to 35.
Last but, not least, the price point for the added capability and functionality will come at a lower cost that the EDW V1 appliance.
I hope you are as excited as I am about EDW V2 and it should be noted that the Microsoft PDW software for the V2 appliance will also run on EDW V1 hardware.
For more information about HP's BI capabilities, please see: