Celerity Adds 'OEM' to Microsoft Silver Competencies

Celerity has once again further strengthened its Microsoft offerings by adding OEM to its ever growing list of Silver Competencies, having already achieved Hosting, Server Platform and Midmarket Solutions Provider competencies.

 

OEM Competency

Partners with the Microsoft OEM competency have achieved a competitive edge with increased customer visibility in the marketplace. Through achieving the OEM competency, Celerity has demonstrated its ability to design and deliver quality hardware solutions optimised for Microsoft Software.

                                                                                  

As a Microsoft Partner it is essential that Celerity is able to deploy, recommend, sell and service the latest Microsoft technologies. By attaining a Microsoft Silver Competency it enables Celerity to set itself apart from the competition by demonstrating a specific, proven skill set to customers.

With the rapid rate of technology change in the marketplace, new product versions open new and exciting solution possibilities. By training and certifying staff on these latest products and technologies both from a technical and sales and marketing perspective, Celerity is able to gain a competitive advantage and is able to offer clients the most relevant solutions in market today.
 

To read this article on Celerity Limited's website click here

When is a Virtual Tape Library not a VTL?

A virtual tape library, or VTL, presents storage as a tape library with tape drives for use with backup software. As most backup software works with tape libraries this means a VTL device will work with all the major software backup applications. More recent VTL’s have introduced deduplication technology for eliminating duplicate copies of repeating data so multiple copies of the same data are only stored once.

So ...  when is a VTL not a VTL?

Well when it is configured with a File System Interface (FSI). This makes it appear to the backup software as Network attached storage (NAS). The following article details my experience configuring IBMs ProtecTIER Deduplication Gateway 7650G with FSI with some of the things you will not find in the IBM Redbooks.

 

The ProtecTIER Filesystem Interface is a relatively new feature, more commonly used on smaller appliance models and not often configured on gateway models where the greater throughput performance of the VTL is more commonly deployed. The 7650G is configured and shipped as either a VTL (with FC adapter cards) or FS (with NICs) but not both. This is something that has caught a few people out in my experience and is not easily rectified once the system is out in the field.

 

 

Unlike the appliance the storage is not integrated with the gateway model so connectivity to some backend SAN storage is required. It is recommended that SATA disks use RAID6 (6+2) and Fibre Channel disks use RAID5 (4+P to 8+P) for User data storage LUN’s and then RAID 10 for metadata LUNs (4+4 to 8+8). It is highly recommended that Fibre Channel disks be used for metadata LUNs as this stores the Hyperfactor index.

 

There is a sizing tool to assist with the calculations for the number of spindles, raid type, disk capacity and performance; this also requires customer information about data change rate, retention period as well as the size and frequency of backups. The sizing information is critical as you will be prompted to input these values when it comes to configuring the storage repository.

 

Network performance in terms of bandwidth between the ProtecTIER and the backup server is an important consideration when configuring the network port. On the ProtecTIER side the NICs come with options of either 2 x 4port GbE adapters or 2x 2port 10GbE. The ports can be bonded or linked together at a logical level to increase bandwidth over that of a single adapter. Round robin is the most straight forward option to configure as this requires no special configuration of the networks switch infrastructure and common sense dictates it is best to link ports across adapters to remove single points of failure. I discussed this with an IBM ProtecTIER consultant and was told this was suitable for performance < 500Mbps. For higher throughput speeds 802.3ad link aggregation is also supported on ProtecTIER but this must also be configured on the network switches and remember that whichever option you choose your backup server should also be configured with the same number of NICs otherwise a potential bottleneck could exist at that end.

 

Configuration of the shares and directories is much simpler than the ProtecTIER VTL setup as there is no requirement to configure tapes, drives, etc. we just need to create a share or shares of the required size. This makes it more straightforward to administer and manage than a VTL configuration. It is worth pointing out that using Workgroup authentication is much simpler to implement than configuring Active directory authentication which requires a detailed knowledge of an organisations AD setup.

 

One last point, in order for ProtecTIER’s deduplication to work it requires data to be written in large chunks from backup software application. You can present ProtecTIER file shares as NAS shares to a file sever but it will not deduplicate, which is the whole point of implementing it in the first place.

If it is a NAS solution you are looking for then a NAS storage implementation is best suited.

 

Should you require any further information regarding Virtual Tape Library please do not hesitate to contact us.

 

Neil Hulme, Technical Consultant, Celerity Limited

 

To read this article on Celerity Limited's website please click here

 

 

- See more at: http://www.celerity-uk.com/news/251/when-is-a-virtual-tape-library-not-a-vtl#sthash.CczWluIY.dpuf

IBM Announce Sale of x86 Server Business to Lenovo

As many of you will have seen, IBM announced, on the 23rd January 2014, the proposed sale of its x86 server business to Lenovo. This is subject to regulatory approval.

 

The threat of sale has been hanging over the x86 business since it was first reported nearly twelve months ago; there is now clarity that this will happen with Lenovo confirmed as the purchaser.

Peter Reakes, Sales Director at Celerity commented “This is positive news for both Celerity and more importantly for our customers.

 

“Lenovo has a proven track record of working with IBM including the purchase of their Desktop business some nine years ago. Since then Lenovo has invested in the products and taken the brand to ‘No.1’ worldwide. This should give us all confidence that our investment in System x and PureFlex technologies will be safeguarded moving forward.

 

"There are still a number of areas which need clarification, not least how the inclusion of PureFlex in the sale will affect the overall Pure portfolio. Celerity will keep their customers informed as the answers become clearer”, added Peter.

 

IBM will continue to invest in their System z brand, power technologies and storage and has also announced a major investment in the 'Watson' technology.

 

Your Celerity account manager will keep you abreast of developments in person but should you have any other questions or concerns please do not hesitate to contact us.

 

To view the full IBM announcement please click here.

 

 

 To view this article on Celerity's website click here

What's New in SQL?

SQL Server is Microsoft’s flagship relational database platform. This article looks at some of the new and enhanced features that will be included in the new release due out in the next version, namely SQL Server 2014.

 

Below are what I consider to be the top 5 features coming to a SQL Server near you sometime in 2014:-

 

1. Cache frequently used data on SSD

Officially known as ‘Buffer Pool Extension’. This functionality allows the integration of solid-state drives (SSDs) into SQL Server 2014 to use as a non-volatile random access memory (NvRAM) extension. In non-techie speak this means that SQL Server 2014 will be able to use the SSDs to cache frequently used data on fast storage that otherwise would have to be retrieved from the slower disks. This feature will work with clusters too as each cluster node can have its own SSDs.

 

My current thinking is that this will probably be version specific; Microsoft uses maximum memory capacity as one of the differentiators between their Standard and Enterprise versions and since this will allow you to extend the amount of available memory it stands to reason that it is probably going to be an Enterprise only feature.

2. AlwaysOn Availability Groups

AGs have been around for a while now, and are a great way of scaling out the read operations on a database system. The fairly significant limitation with AlwaysOn in SQL 2012 was that should your primary fail, or should your cluster lost its quorum, then it would automatically take down the readable secondaries too.

 

With SQL Server 2014, this will no longer be an issue; the secondaries will remain online even if the primary is no longer available. There are obviously a few caveats to this, the main one being that you will not be able to query the AG listener and will have to connect to the replica’s server name instead. This is a fairly easy thing to work around if you create individual DNS records for the readable replicas, but it is something that you will have to keep in mind. Additionally, you can now have up to 8 secondary AlwaysOn servers (previously 4), albeit with the increased cost in Enterprise licenses too.

 

3. Failover Cluster Support Enhancements

In the olden days (i.e. prior to SQL Server 2014), if you had a SQL Server Cluster then you had to accept the rule that only one node could access a volume at a time. This node effectively ‘owned’ the volume and no other node was, therefore, allowed to read from, write to, or for that matter see it.

 

However, from SQL Server 2014, you will be able to take advantage of using Clustered Shared Volumes in your cluster; this gives you a much greater flexibility and allows multiple cluster nodes to access the same storage volume at the same time. You will no doubt be aware that this technology been available in Windows and Hyper-V for a while now and it’s good to see the same functionality coming to SQL Server too. The biggest gain that I can see with this feature is that should one node of your SQL cluster lose connectivity to its storage, it should still be able to read and write data over the network to the storage a via a different nodes connection. Obviously this would be slower than if it did it directly, but it is still a quantum leap over and above the functionality we had with traditional clustering.

 

4. Specialised in-memory OLTP tables

If your organisation has a truly large dataset (the term ‘big data’ springs to mind), then this feature may be of extreme interest to you.

 

Hekaton is apparently Greek for a hundred, which is the performance improvement goal that Microsoft set out to achieve when they started with this new technology. Hekaton is best suited for storing ‘hot’ (heavily used) tables in memory effectively removing the need for latches and locking to occur, which in turn should then result in a significant overall performance increase.

Obviously with this sort of potential improvement, it is not going to be as simple as enabling the feature and away you go. There will be structural changes required to implement it properly. For instance, Hekaton does not support the use identity columns, which means that you may need to consider changing your primary clustered key to a GUID instead.

 

Also, you will hopefully have already considered that keeping specific tables resident in memory will mean that you have less of it available for the rest of the system to use. This may not be a problem if you have a monster server or two with terabytes of RAM available, but for the less ‘Olympian’ administrators out there you should at least look at whether you need to max out the memory capacity for your servers before turning this feature on. You definitely do not want to run out of available memory with this feature.

 

5. The Cloud

Although already mentioned previously in passing, Microsoft is also investing heavily in their cloud offering Windows Azure. New features and functionality in SQL Server 2014 will allow you create hybrid database environment that allows for easy wizard based movement of databases between on premise and Azure in both directions.

 

This functionality also extends to AlwaysOn groups, sending backups directly to Azure storage and for that matter it will even be possible to create an on premise database, with the data and log files held in Azure. Whilst I am the first to say that I am not exactly sure as to why you would want to do this last one, I sure that there are some scenarios whereby this would be a desirable thing.

 

Final Thoughts

As a former SQL DBA, used to managing multi-terabyte databases, I can see a number of potentially significant performance improvements that will be coming in SQL Server 2014. It’s good to know that whilst Microsoft are advancing with their Cloud vision, they are also still investing heavily in improving their on premise versions too, providing administrators with new features to allow them to implement a hybrid on premise/cloud environment when you’re ready.

 

Whilst there is no pricing information for SQL Server 2014 available from Microsoft at present, I think it is reasonably safe to say that it will not be any cheaper than you are currently paying for SQL Server licensing.

 

If you have the time, I would suggest that you download a copy of the latest preview and have a play. Microsoft have stated that if you have a database that currently runs on SQL Server 2012 you should have no problems running it on SQL Server 2014.  I would take that with a pinch of salt as there are always going to be exceptions, but give it a go and see what performance gains are there to be had out of the box. I think you will be suitable impressed.

 

Please note: The information in this article is based on the latest community technology preview and the feature set included in the final build is subject to change until SQL Server 2014 goes to RTM.

 

Should you require any further information regarding SQL or to discuss your requirements please do not hesitate to contact us.

 

Mike Metcalf, Technical Consultant, Celerity Limited

Contact
 

 To View this article on Celerity Limited's Website please click here or visit www.celerity-uk.com

What is Scrum?

Historically a large proportion of companies have used Prince2 or a derivative for their standard project methodology. However, over recent years the move of companies to a more dynamic and flexible approach to projects has increased. There are a variety of these new approaches e.g. scrum, lean etc., which generally are more about being adaptive.

 

Scrum is a framework for managing projects that is focused on a flexible holistic strategy where a team works as a unit to reach a common goal as opposed to a "traditional, sequential approach". This method was originally designed for use within application and product development, but can also be used in other areas. Scrum enables the creation of self-organising teams by encouraging co-location of all team members, and verbal communication among all team members and disciplines in the project.

 

A key principle of Scrum is its recognition that during a project the customers can change their minds about what they want and need (often called “requirements churn”), and that unpredicted challenges cannot be easily addressed in a traditional predictive or planned manner. The focus is on maximizing the team's ability to deliver quickly and respond to emerging requirements.

 

There are only three roles involved in this methodology:

 

• Product Owner – this is the person responsible for the Product Backlog i.e. contents, priorities and availability

• ScrumMaster – this is the equivalent role to a Project Manager from a responsibility prospective, but has major differences in that this is more of the facilitation role rather than a management role

• The Team – is responsible for the process of each iteration during the project and the project itself

 

Scrum projects generally start with a vision, which might be vague at first, but will then become clearer as the project moves forward. The Product Owner is responsible for delivering the vision that maximises the ROI and formulates the Product Backlog.

 

The Product Backlog is a list of functional and non-functional requirements for the solution delivery by the project. The Product Backlog is never complete as it evolves from an initial estimate of the requirements through management identifying changes to make the product improvements. The items are then prioritised so that the items can then be grouped into proposed releases.

 

All work is done in Sprints, which are manageable chunks of work (maybe 2-30 days) where the objective is to get a ship-ready state.

 

Sprints generally begin with a planning meeting and end with a review meeting.

 

A Sprint Backlog is a list of tasks to be completed during the sprint and assigned to individuals. These are the actual decomposed items from the product backlog. This list must be agreed to by the entire team prior to the sprint actually beginning.

 

The Scrum is a short regular meeting, which is quite often daily, where the usual brief is to answer the following three questions:

 

• What have you done on this project since the last Scrum meeting?

• What do you plan to do on this project between now and the next Scrum?

• What impediments stand in the way of you achieving your commitments to the project?

 

A Burndown Chart is a monitoring tool, which shows the amount of work remaining across time and is a good way of visualising the correlation between the amount of work remaining at any point and the progress of the Team(s) in reducing this work.

Scrum can be implemented through a wide range of tools. Many companies use universal tools, such as spreadsheets to build and maintain artefact’s such as the sprint backlog. There are also open-source and proprietary packages dedicated to management of products under the Scrum process. Other organisations implement Scrum without the use of any tools, and maintain their artefact’s in hard-copy forms such as paper, whiteboards, and sticky notes.

 

Should you require any further information on Scrum please do not hesitate to contact us.

 

Kevan Dix, Project Manager, Celerity Limited

 To view this article on the Celerity Limited website, please click here

IBM Storwize - 'Taking a Technical View'

When IBM launched their all new, virtualised storage “Storwize” portfolio a few years back it consisted of a single product - the V7000. Today this is still the top-of-range model but the portfolio has expanded to include a couple of other systems.

 

The second model launched was the V3700. Like all in the range, this system runs the same SVC software as its V7000 big brother and so benefits from the same highly-intuitive GUI and ease of management, but has less hardware grunt inside and so sacrifices scalability and a couple of features. Scalability on the V3700 is considerably reduced as it only supports half the internal storage per controller and controllers cannot be clustered. Additionally, external storage virtualisation is lost meaning that the V3700 cannot manage external storage systems as if they were its own storage.

 

This is where the V5000 comes in. It fills in the gap between the top-of-range V7000 and the baby V3700. Using more powerful processing power and cache in the controllers’ two nodes, more storage can be attached to the V5000 and two V5000s can be clustered together. More key, however, is that SVC style external storage virtualisation is supported meaning that, like the V7000, it can manage external storage. This brings one of the key benefits of the SVC and V7000 code to the market at an even lower price point.
 

 

The V5000 offers a range of connectivity options including 10Gb iSCSI, FCoE or 8Gb FCP connectivity. As illustrated above, the appropriate I/O daughter card is installed meaning that while the V5000 does not offer the connection versatility of its bigger brother (as the V7000 natively offers all these protocols simultaneously) it can be configured to be suitable for either type of deployment.

A couple of other features are amiss such as compression, and extra cost licences are required to activate some other features that are included in the V7000 base code, but the V5000 is still a very powerful storage system that delivers everything many SMBs require at a price point making the technology even more competitive and affordable for more businesses.

 

The table below summarises some of the difference between the products in the Storwize family:

 

Per system so clustered models offer multiple of this value

Should you require further information regarding the IBM Storwize V5000 or have any storage requirements please do not hesitate to contact your Celerity Representative.

 

Edward Yates - Technical Consultant - Celerity Limited

 

 To view this article on Celerity Limited website click here

Benefits of Virtualising your Storage

The main benefit of virtualising your server estate – is to maximise the utilisation of the hardware and reduce the physical server footprint - has been known and proven to I.T. professionals for many years and adopted in some way by most.

 

In addition to massively increasing the efficiency of the server estate and reducing the procurement and running costs of each (virtual) server, new functionality becomes available that would not otherwise be possible. This includes things like migrating the server from one piece of hardware to another, taking instantaneous backups of the server as a whole and moving it from one storage system to another, all without any disruption.

 

But typically many organisations stop there, yet most of the benefits of virtualisation can be brought to the storage too. By virtualising the storage platform you gain improved functionality and more advanced storage features. Not only does storage virtualisation enable more common features such as instantaneous point-in-time snapshots for back-up for example, but can also bring new features such as:

• Instantaneous cloning of volumes to allow testing and development of production data without any impact on the live services
• Manual and automated migration of volumes between different tiers of storage within and sometimes across storage systems
• New ways of protecting data and restoring system resilience after hardware failures allowing for faster rebuilds while reducing or eliminating the performance impact on the system
• Advanced functionality such as compression or de-duplication of tier 1 production data can be done at the storage level and is transparent to the applications and servers
• Offering both NAS and SAN connectivity protocols from a single, unified storage system to streamline system management

 

Not only does storage virtualisation bring these features to your environment along with a performance boost, but often does so in a smaller footprint which can dramatically reduce the total cost of ownership.

 

So we know what storage virtualisation can bring, but how is it achieved? Typically virtualisation of storage “achieves location independence by abstracting the physical location of the data”, or creates a new virtualisation layer (or pool) of storage and all the magic happens at that layer, rather than at the actual physical disk block level.
 

As oppose to data volumes residing on dedicated RAID arrays which need specific characteristics to meet the application’s demands, volumes are created at this storage pool layer which can be made up of multiple arrays meaning more performance and greater disk utilisation thanks to the large number of disks in the pool, yet offering ease of storage design and management.

 

Should you require more information on Virtualising your Storage, please do not hesitate to contact your Celerity Representative.

 

 

Edward Yates - Technical Consultant - Celerity Limited

To read this article on our webpage please click here