IT managers know that the procurement, development and deployment of Information Technology hardware and software is a complex process. Part of this complexity is by nature but the other part is by design. A dirty secret in the industry is that when a customer does not understand his options, he can easily be led to spend more money than he should. To protect oneself, a user has a responsibility to research the latest innovations in the technology area of interest. This is easier said than done since vendors tend to push as much hype, complexity and confusion into the market as they can.
Caringo takes a fresh approach by helping users solve one piece of this puzzle. Our flagship product, CAStor, stores content, and our corporate mission is not only to provide the best technology available, but to change the very economics of content storage. Caringo seeks to peel away the layers of complexity surrounding the content storage business, which in most corporations is 60% or more of total data. In fact, according to The InfoPro, average installed capacity in Fortune 1000 companies has jumped from 198 Terabytes in 2005 to 680 Terabytes in 2006. That number appears to be doubling every 10 months! This document provides some guidelines to use in thinking about and procuring content storage systems.
Information Lifecycle Management or Installation Lifelong Mortgage.
Information Lifecycle Management (ILM) vendors would like you to store all your data on the most expensive and fastest disks they can sell you. Once stored, they assure you that you can take your time deciding where and when to move the data to a more cost effective platform. The first flaw in this argument is that you should use the most expensive and fastest platform for everything upfront. Data has different requirements depending on its source and its use in your business and frankly, you should know these details before you store data the first time. If you store your data on the right platform the first time you store it, you save time, money and the headache of wondering when and where to move the data later. ILM vendors would also have you purchase the rather costly software that helps you decide when and where to move the data. Moreover, they would like you to purchase three different types of hardware to make sure that you have plenty of options. Complex? You bet. The truth is that a well run IT shop knows the types of data it is storing before it is stored and a well run IT shop is smart enough to store the right data on the right platform the first time. It's easy to get intimidated by vendors who tell you they know more than you do. But no one knows more about your shop and your data than your IT personel.
The numbers don't equate.
High speed disk can run as high as $60 per gigabyte. content storage (hardware and software) can run as low as $2 per gigabyte. So let's assume you are storing 100 gigabytes of content (a rather small amount of data these days) per week. You might allocate an additional 20 gigabytes of storage (it's always a good idea to allocate 20% extra for growth) at $60 per gig. That would be around 3 Terabytes of storage every 6 months costing $180,000 for the high speed disk. We'll throw in floor space, power consumption and software into that $60 per gig price (even though these extras will probably cost more).
As a comparison, using Caringo's CAStor content addressed storage software running on commodity hardware, you can configure a 6 Terabyte system. (This is the raw space.) If you replicate all your content, the useable space is 3 Terabyte) using six x86 nodes with two 500 Gig drives in each node for a total of 6 Terabytes per node. Add the hardware and software together and the grand total would be about $15,000 for 3 Terabytes of usable content storage that not only stores your content reliably, but also makes it available online and in real time. That's a greater then 10 time reduction in cost for storing your content. Add in floor space and power consumption and you are still far below the cost of expensive disk.
An additional cost factor to consider is that CAStor is self-managing storage. Adding additional capacity requires no provisioning or management, simply plug in a new node, add a CAStor memory USB key and power it on. The self managing behavior of CAStor significantly reduces the overhead on your IT personal allowing them to add value to your business. This reduction in TCO provides an additional factor of up to 10 times the reduction in cost over the life of your content storage. That's not even considering what you might save due to the hardware agnostic nature of CAStor preventing vendor lock in that inflates costs.
Numbers don't lie.
When procuring systems, there are a number of fees one must take into consideration. Hardware and software acquisition costs, media costs, hardware and software maintenance costs, floor space and power. The percentage of these specifics vary per type of storage. Generally, the adjusted component cost percentages are as follows:
| TYPE |
HW |
SW |
Media |
HW Maint |
SW Maint |
Floor Spc |
Power |
| Tape |
75% |
10% |
3% |
5% |
5% |
2% |
1% |
| DVD |
40% |
13% |
27% |
7% |
7% |
5% |
1% |
| Disk |
37% |
45% |
0 |
5% |
7% |
1% |
5% |
| Typical System Component Costs (%) |
There are clearly a number of policy and corporate decisions that will drive your decision on the storage of content. The interesting new facts show that tape is no longer the lowest price option. In fact, if your business requires on-line access and application of policy constraints (such as reliably deleting content after a specific time period), a disk to disk content storage solution becomes compelling. In the interest of comparisons, we have added two disk to disk content storage solutions below. One is a Market Leading Archive Vendor (ML-AV) and the other, Caringo's CAStor.
For a 10 Terabyte usable archive for content storage running over 3 years, the numbers look something like this.
| TYPE |
HW |
SW |
Media |
HW Maint |
SW Maint |
Floor Spc |
Power |
TOTAL |
| Tape |
116,000 |
10,000 |
2,250 |
4,000 |
4,500 |
1,600 |
850 |
$139,200 |
| DVD |
36,500 |
12,000 |
25,000 |
6,500 |
6,000 |
4,000 |
800 |
$90,800 |
| ML-AV* |
110,000 |
135,000 |
0 |
13,000 |
20,000 |
2,500 |
14,000 |
$294,500 |
| CAStor |
30,000 |
20,000 |
0 |
3,300 |
12,000 |
2,000 |
10,000 |
$77,300 |
Typical Content Storage Costs ($) 10 Terabyte usable archive running 3+ years |
| Note: |
Prices are taken from industry list price lists. We all know that a good negotiator can get better pricing.
|
* Market leading archive vendor (disk to disk)
Data Protection Options.
There are many other key elements to keep in mind. Tape is an interesting medium. It can be lifted and stored offsite (although it has been known to fall off trucks). Often, weather and storage conditions can decay the data stored on tape, and after 10 years should be moved to new tape cartridges. For many years, tape was the defacto standard for long term storage, but today, the many risks and the fact that data that should (by policy) be destroyed, often continues to exist on some tape in a cavern somewhere that may or may not surface when the need arises or worse may surface unexpectedly during legal discovery. Moreover, it takes time to find a tape, mount it and search for the data. And as new technologies develop, it is sometimes hard to find the software that created it and a device on which to play an old tape. Worse still valuable digital assets are not available on line to be leveraged.
Optical technology also has challenges. While claims have been made that the life expectancy of a disc is 50 years, it is expensive, unreliable, hard to delete and easily broken or damaged.
Disk is by far the preferred technology for long-term storage today and as the price decreases more and more, there is little reason not to use it. So let's assume that disk is affordable and preferable. What other elements should you watch out for?
Many hardware vendors require you to purchase a large storage platform that you will not fill for many years. There are only two reasons for this. First off, it makes a nice purchase order for the vendor and second, most of these systems are not very scalable.
Data Protection That Makes Good Business Sense.
CAStor provides a number of unique and cost effective advantages.
It runs on commodity hardware and is massively scalable. This means you can begin with a small cluster - say 2 or 3 Terabytes - for a very small price. As your storage needs grow, you can add nodes to the cluster on the fly. There are essentially no limits to CAStor scalability. Keep adding nodes and your performance and storage keeps getting better and bigger. This actually has another amazing advantage. As commodity hardware gets faster and better, you can literally upgrade the hardware in your cluster just by bringing up new nodes with more modern hardware. This seamless upgradeability in itself is a great cost and time saver. CAStor automatically configures the new node so there is no Sys admin work to worry about. No technical or service call required. In this same manner, you can remove older hardware from the system. Simply 'retire' an older node, that node disperses its content across the other nodes and when it is done, shreds the contents of its disks so you can safely re-deploy or dispose of it.
CAStor also heals itself when a disk goes bad. The recovery can be faster than RAID (depending on the size of the cluster) and your data remains available during recovery. In fact, you can even leave the bad disk in the node until it is convenient to replace it. We call this 'weed at your leisure.' It's another time and repair call savings unique to CAStor.
Be careful not to fall into the vendor trap when looking at content storage. Remember that technology changes often, so the choices available today, might not have been available last year. Faster and better systems are usually brought to market by start up companies simply because the long-term players have so much baggage to contend with. Industry standards, scalability, commodity hardware, ease of maintenance, cost-effective management and long-term viability are all areas you should examine fully when preparing a content storage procurement. The selection of the best technology at a price point that makes good business sense is the only one you should make. And of course, Caringo would be pleased to assist you in your search for the most cost effective, high performance, scalable and easy to manage system for your corporate needs.
Please contact us at sales@caringo.com for more information.
|
|