Alok Sinha, AVP General Business Sales and IT Solutions Head, Huawei Enterprise India.
There was a time when your application required more storage space, you would simply add more disks to your server (IDE/SATA/SAS) and that would solve the problem. Six months later, the data asset size increased and therefore, the internal disk based storage was no more sufficient. However, the application still required that additional disks be available as native local disks, and not as mounted disks on a remote server. This was achieved by putting in a storage solution that was connected over fiber channels, collectively known as Fiber Channel based Storage Area Network.
Another six months pass and yet more space is required. But now, there's a new problem--the local datacenter is running out of space. The new disk-arrays need to be hosted at another geographical location. Unfortunately, the IT budgets are a little tight this year. It's time to choose the iSCSI option. In this option, any generic ethernet device can be used (unlike the special device requirement of SAN-FC), to connect to this new array of disks. In layman terms, iSCSI allows two servers to negotiate SCSI commands over a standard Ethernet network. The methodology is known as SAN-iSCSI and allows the network to be peaked upto 1G/10G. Due to the fact that native IP routing is used, your current infrastructure — network switches/routers do not need any modification.
Some more time passes, and we need even more space. This time, we need something that has the ability to provide high speed access and very high reliability for mission-critical applications. In comes FC over Ethernet (FCoE) and 10G Ethernet to create additional disk arrays, hosted at a different locations, while continuing to give high speed SAN for your applications. Due to the fact that IP protocol is not used, there are no overheads and the performance in speed is un-paralleled.
With the passage of more time, another application gets deployed, which requires server data to be picked up from a partner — over the net — perhaps using an SMB/CIFS. You rush to deploy a NAS solution that allows you to mount file systems over your existing environment.
Finally, let's throw a bit for complexity into the picture--we need to take backups, not only at the file-level, but also at the disk block levels
Enter Unified Storage
In the above scenario, over a period of time, you're left with lots of different types of hardware, drivers, software to work on, manage and support. None of the above technologies are in lieu of one another, but rather they all co-exist in the same world. Most CIOs who have handled the above scenario would know too well the pain points of owning such a storage infrastructure. This is where unified storage comes to rescue. As the name suggests, a unified storage device is a single integrated storage infrastructure that simultaneously supports the following:
- Fibre Channel (FC)
- Fibre-Channel-over-Ethernet (FCoE)
- IP Storage Area Networks (iSCSI) and
- Network Attached Storage (NAS) data protocols.
With all storage technologies in one box, managing the complex storage infrastructure scenario above suddenly becomes easy to manage. It's a technology that's seriously being considered by a lot of organizations.
Benefits of Unified Storage
Deploying a unified storage solution can result in the following benefits for an organization:
- Better utilization of existing storage resources: Most efficient IT managers, who pride in optimal utilization of their resources will reveal to you in private that they are certain that some disk is hidden somewhere, which may not be on his radar and thus he would not be always sure that all storage units are indeed being used most optimally. Building a NUS allows you to map, manage and optimally use all the resources, that you can possibly have.
- Non-disruptive data migrations: Ask any IT specialist, his most challenging time, and it is certain that he will recount how migrating from one system to another, without having the flexibility of a down time, was one of his worst work-place nightmares. More so, if it involved storage and if it involved non-standard device migration. Most NUS provides this as a simple few click options to manage non-disruptive migration, conveniently.
- Standardized approach to storage management: IT skills are at a premium, keeping your team's skill updated in all the different technology and product management requires significant continuing investment. This can be abstracted by deploying the NUS layer, so that all the configuration and management of varied tools are deployed using common GUI. Complexity of RAID and volume and LUN management are no more the domains of geeks alone.
- Consolidated licensing model: Different hardware and software means, managing different licensing models and thus different costs. One has to be extra careful when you deploy the NUS infrastructure. One simple low cost box can save you millions of dollars in licensing costs alone. This is one of the most direct and tangible return that you can leverage, to support your decision to move into NUS world.
- Serving heterogeneous customer needs: As a service provider and data center manager, you may not always have control on the specific application that your customer (both internal and external) demand. Thus, it makes logical sense to have the ability to offer services that transcend technology and physical boundaries. There is no logical reason, not to deploy NUS in this kind of environment.
When Not to Use Unified Storage
So we listed most of the good reasons on this planet on why you should choose a Network Unified storage solution. However, we also came across a few situations where choosing NUS may not be the best option. It's not the panacea to all your storage woes, nor is it a replacement for all other storage technologies. So you are going to require other storage technologies as well. Here are a few scenarios where choosing a NUS may not be suitable.
- Single vendor dependence: One strong school of thought, is the uncanny resistance of most CIOs to choose a single vendor. While the debate over single or multi-vendor solution rages from time immemorial, it may be of interest to note, that about 10% CIOs are not comfortable putting all eggs in the same basket.
- A specialist is a specialist: The network unified storage is like an all-rounder, capable of doing several things (read devices and protocols) and excelling at them. However, like any good team, you need all-rounders and you need specialists. Many CIOs are still not comfortable diverting mission-critical applications on unified storage. They still prefer and for good reason, to depend on the specialist technology like — SAN-FCoE or SAN-FC to deliver high performance solution in storage.
- Not going to use all the components: If your needs for the next 36 months does not include the need to go up all the components, there is a need for you to do the cost analysis of the NUS vs SAN/ NAS. Most modern application have learnt to get over the dependence and the need for local disks, they are comfor-table over network storage (NAS). Don't spend money in technology that you do not intend to use, even though the cards may be modular and choices many — you are still paying for a technology that you may never use.
Features to Consider in a NUS
With a plethora of vendors offering unified storage solutions, choosing the right partner will not be easy or fast. One will need to track down and compare the features that meet your needs. Some of the essential features you should consider are as follows:
- Power down backup: Irrespective of the best of power management systems that you can deploy, there will be one day that the power supply would fail. On this un-fateful day, you would be shocked to see, all files and file systems reporting severe corruption errors. This can be totally avoided if your NUS has the ability for soft power down and power failure protection.
- Redundancy in FC channels, controllers, power supplies, fans, UPSs and disk enclosures, will allow you to offer a reliable solution.
- Disk types supported: Check how many different types of disks does the NUS support.
- Key features: Snapshotting, de-duplication, mirroring, backup abilities, and LUN copy
- File Systems Support: NFS V2 or V3, CIFS, iSCSI, FCP, Fabric, FTP, HTTP, NDMP
- Service and support: Must have 24x7 support system to deliver your customer's need for full time access.
- Access control: NIS, Active Directory, LDAP
In summary, unified storage or NUS is the next generation and the logical step of maturity in storage. Its potential to consolidate and integrate the wide spectrum of technologies and devices is legendary. However, in the business case to deploy NUS, you would also want to consider the real needs and benefits of NUS over existing specialists of SAN/NAS storage.