Archive for the Tag Dynamic Infrastructure
(formerly known as the IBM System Storage Symposium.)
If you sign up early you get access to both the Storage University and the System X and Blade Center University. They are at the same time and you can walk from one to other as needed.
There is a ton of information about IBM Storage at this event, if you have never been before, you should atleast take a look at the amount of material that will be covered.
IF your are interested in presenting or would like to know more feel free to reach out to me and I can put you in contact with our staff.
Posted by richswain in Oct 26, 2009, under Event
Event at the NYC Netapp office. Gain perspective on the business and architectural considerations required to support an end-to-end,
virtualized data center. Learn more about how the IBM N Series Storage uniquely compliments vSphere environments. Understand how the proven technologies such as deduplication, thin provisioning and snapshot can significantly reduce the amount of storage needed in a virtual environment while enhancing performance and streamlining administration.
Sponsered by IBM, Netapp, VMware and Brocade
The Virtualized Dynamic Data Center
Wednesday, October 28, 2009
3:30 – 5:30 Virtualized Dynamic Data Center Presentation
Breakout Sessions: Cloud Computing, Brocade,
IBM System X
Wine Tasting Reception
100 Park Avenue, 13th Floor, New York, New York -
Posted by richswain in Oct 23, 2009, under News
IBM N series Release News 4 of 4
TGIF everyone! Our final installment of the IBM N series release news. I have updated you on the new generation PAM II Cards, SnapManger for Hyper V and today we will discus 10GB Ethernet/FCoE.
Consolidation is a great topic you will find lots of bloggers talking about and we mainly talk about either server consolidation with virtualization or storage consolidation with deduplication or cloning technologies. But now we have the technology to consolidate your network topology into larger pipes using 10GB ethernet.
This idea of moving from a slower network speed like 10 MB to 100MB or 100mb to 1GB has been all based on faster performance and building a bigger pipe to applications. Now that we can push 10GB there is a different idea behind the movement.
The movement comes on the heels of virtualization and ‘Doing more with less’ mentality. No longer are we looking for just faster speed, but we are looking at the cost that we can save by consolidating all of the 1GB links down into a single 10GB link.
I still think most people will recommend two 10GB links to help with high availability but why not use 1 10GB port as a primary and then a few 1GB links teamed (Vifed together) as a backup or failover. This will allow you to utilize the speed and performance with the hit of having to use two 10GB ports on your switch.
Just like the 1GB switches when they arrived on the data center scene, 10GB switches are still expensive. But if you look at history, the cost per port will start to come down after the ‘newness’ of it wears off and more and more people start adapting it as their standard.
IBM N series now supports the 10GB ethernet card with dual ports. This allows you to connect to two different switches or create a 20GB vif. The setup of the 10GB is the same as the 1 GB ehternet ports you have been using for years.
IBM N series is the first storage product (and only as of now) to support FCoE. The FCoE cards do come in two different flavors, fibre or copper ports. They are both setup as target cards and use a PCIe slot.
Posted by richswain in Oct 22, 2009, under News
IBM N series Release News 3 of 4
The past two days I have been writing about new technologies IBM released on October 20th. There was plenty of hardware to help IBM customers to get more bang for the buck but there was also an addition the SnapManager suite; SnapManager for Hyper-V.
One of the biggest chances in the data center has been the server virtualization of the data center. This new idea of consolidating the server farm from multiple 1 U servers to a larger server hosting multiple virtual machines posed new challenges. Backup, restores, IO bottle necks; all new things we had to work through.
For many years VMWare was the main player in the virtual server game. As the virtual movement caught more traction with customers, other vendors started coming out with their own virtual products. Microsoft released their virtual product codename Viridian, with their Windows 2008 OS and named it Hyper-V.
IBM SnapManager suite allows customers to take a consistent backups of databases and Hyper-V is no different. Built on snapshot technology, SMHV takes snapshots of the VM. Some of the terms used with Hyper –V is different than VMWare but the idea is pretty much the same.
SMHV uses a group of objects called datasets. These datasets can have certain retention policies set on the dataset. The dataset is backed up collectively but can be restored individually.
There is also Snapmirror integration that helps mirror the datasets to another N series set of disks. Reporting is also a feature that is included.
If you have some time to review the SMHV product, there are a few links you should take a look at:
Posted by richswain in Oct 21, 2009, under News
IBM N series Release News 2 of 4
As I noted in yesterday’s blog update, IBM has released a wealth of new hardware and software on the N series platform. This release includes the first IBM storage platform to support FCoE and 10Gb Ethernet, a new disk shelf that is 30% more capacity than older technologies in a 4 U footprint, an additional SnapManger product called SnapManager for Hyper-V and the next generation Performance Acceleration Module. The technology behind the N series is geared to help our clients meet the new paradigm shift of doing more with their storage with higher availability, quicker recovery and less cost.
Today, I wanted to update you on the new generation of Performance Acceleration Module or PAM II for short. This second generation technology is built on the predecessor of only 16GB of memory. The idea is to increase the amount of read cache by using flash-based technology instead of adding more spindles to the entire system.
Traditionally we have needed to add more disks into the spindle count to increase our IOP throughput and lower our response time from a disk subsystem. Some have changed the disk type to a solid state disk which allows faster response and higher IO throughput for the data that is only on that chip. This can get expensive for larger data sets that still require ‘Tier 1’ throughput and it only benefits the data that is stored on the chips.
A comparison study (link) was done to see what the difference was between adding more drives or adding PAM II cards. A N6060 system with 84 disks were tested with an OLTP workload which is typically both random and sequential IO. It has been described as the best method to really test performance by not only IBM but most other storage vendors. The result was it took 140 more disks to the same throughput of adding 1 TB of PAM II cache. The cost for the PAM II cards was 50% cheaper than the purchase of 140 more disks. The other cost savings was in additional power and cooling and a total savings of 30 U in the rack.
The PAM II cards are SLC NAND Flash memory instead of the older SDRAM and comes in both 256GB and 512GB sizes depending on the N series system. A total of 2TB of cache can be added to the system for improving performance. They do require a PCI-E slot so only the N 6000 and the N7000 product systems can take advantage of this new technology.
If you are looking at increasing the performance of your system and want an alternate to spinning up more disk, PAM II might be your answer. PAM II is designed to help you with read heavy workloads like OLTP databases, messaging and virtual infrastructure. If you want more information about PAM II feel free to contact your local IBM Storage rep or a IBM Business Partner.
Posted by richswain in Oct 20, 2009, under News
IBM today announced releases for the N series platform including the first IBM Storage product to offer FCoE support. The N series platform now includes
- Performance Accelerator Module – Second Generation
- SnapManager for Hyper-V 1.0
- Fiber Channel over Ethernet
- 10Gb Ethernet support
- EXN3000 SAS Expansion Disk
Over the next 4 days, I will be writing about these technologies and how they related to the IBM Smarter Planet initiative. With these technologies, you will be able to consolidate your data center infrastructure, use less power and cooling and make your solution more efficient.
With budgets getting tighter and CIOs asking to do more with less, companies are asking IBM to help them achieve this shift in the paradigm. The IT department is being not only to provide more services at a higher availability but some are asked to be revenue centers for the company.
The new Performance Accelerator Module II (aka PAMII) is an upgrade to its smaller brother the PAM I. The new flash memory card can be added to systems (both N6000 and N7000) in 256GB and 512 GB cards. Unlike Solid State Drives that accelerate data stored on the chip, the PAM II effectively extends the buffer cache for all data. By adding memory cache to the system, it reduces the total number of IOPS the disk subsystem has to perform across the entire dataset stored in the system.
Server virtualization has become a hot topic with many companies looking to consolidate the data center footprint. Now IBM is adding to the list of SnapManager software packages support for MS Hyper-V. SMHV is built on the SnapManger product line that cut it’s teeth on Exchange and Oracle many years ago. Much like those products, SMHV allows customers to backup and recover their virtual machines within the storage protection policies set by the backup administrator. Clients that use SMHV will find it easier to rapidly provision and clone their virtual machines and the ability to recover one virtual machine if needed.
Server virtualization has been a hot topic for a few years, and people have started looking at storage virtualization/consolidation within the last year. Something that was left out was the networking in the data center. The introduction of Fibre Channel over Ethernet (FCoE) and 10Gb Ethernet is allowing companies to save more money by consolidating older 1Gb lines into one or two lines. IBM N series is the first storage system to support FCoE hardware. We are very proud to have this honor and as customers look to IBM to provide an FCoE or a 10Gb product, we have a proven product that will help consolidate their network infrastructure.
The new announcement of the EXN3000 disk shelf is part of the IBM Smarter Planet goal. With a total capacity of 48TB raw disk in a 4 U footprint, we can reduce the number of watts consumed (per TB) more than 10% per shelf. The EXN3000 can use either SAS or SATA drives in 12 or 24 drive configurations. This equates to about 30% more dense storage than the EXN2000 or EXN4000 shelves. With a higher density storage footprint, less power and cooling is required for fewer spinning disks.
These technologies allow IBM customers to meet the tightening parameters that are changing the paradigm in IT. IBM is constantly building on the proven N series platform to enhance and simplify our customer’s environment while increasing efficiency with the world’s resources. Come back tomorrow for part 2 of 4 in the week long update on the IBM N series release.