This article can also be found in the Premium Editorial Download "Storage magazine: Tips for unifying storage management."
Download it now to read this article plus other related content.
The IT world couldn't be happier. After all, no sector in the economy lived through a more pronounced riches to rags story than we did. Once again, stocks are up and tech is hot. This good IT economic news leads us down a logical path. Positive stock market activity will likely lead to a renewed--albeit more sober--IPO market in 2004. IPO action will lead to more VC funding, which will in turn lead to more technology startups to solve IT problems.
Startups certainly have a schizophrenic past. They were a welcome addition to the IT mix back in the boom years, and were innovative and driven. They had a certain swagger about them as they readily proclaimed that their technologies would cause a paradigm shift. It all sounded so believable. Then suddenly, well-funded storage startups such as Cereva, Scale8 and Storage Networks were out of business while others such as GiantLoop/Centre Path and Storability Software were forced into alternative business models. IT managers felt burned and deceived and as a result, they banished startups from their shops. In many companies today, startup vendors will be lucky to get a return phone call, let alone an order.
Times have changed. Startups
You may ask yourself--why bother? After all, the world is full of established storage companies such as EMC, Network Appliance (NetApp) and Veritas that offer good products. But large companies become bureaucratic, risk averse and protective. Employees tend to work nine-to-five hours. Startups stick out their necks to be more innovative and execute better than the big guys. In the storage world, this is a good thing as it drives change, enhances competition and provides IT with good alternative solutions to vexing storage problems. Just look at some of the startup activity around:
- Information life cycle management (ILM). Moving storage files around over time will be an industry-wide effort combining hardware, software and services. Companies such as Arkivio, KOM and Seven Ten Storage Software offer products to accelerate this effort.
- Object storage. The Enterprise Storage Group, Milford, MA, and others believe that the biggest driver of storage growth over the next ten years will be unstructured data. EMC created Centera to capture this market. Now startups such as ExaGrid, Permabit and Reference Information Systems are following in EMC's wake.
- Enterprise network-attached storage (NAS). NetApp recognized this growing space by acquiring startup pioneer Spinnaker. Isilon Systems, Panasas and Pillar Data Systems are other startups looking for wins in this space.
- Alternative backup. It's official: Tape backup alone can't address the growth in storage. Creative startups such as Alacritus, Diligent and Sepaton want to fill this gap.
- Storage operations software. How can IT manage distributed heterogeneous storage? With storage operations software. This emerging market has attracted all of the old guard as well as newer companies such as AppIQ, CreekPath and NuView.
Define the ideal solution. Before surfing the Web or picking up the phone, define a solution that would best fit your needs if you were building it yourself. Give yourself a little intellectual freedom here. How would this technology be packaged? What would it need to support and integrate with? What would the user interface look like? Make sure to assess the requirements from an IT and business perspective as this will help you define things like availability needs, service levels, reporting capabilities and pricing. Think long and hard about the metrics you will use to gauge the product's success or failure. This exercise will help you prepare RFIs and RFPs and streamline the sales process for your technology vendors.
Check references. This may seem obvious, but you'd be amazed at how cavalierly many companies approach this process. Vendor references are your chance to assess every aspect of a startup, not just its product. CIOs should demand at least three references from each startup. Remember this is the time to ask everything about the startup: financials, products, skills, processes, etc. The IT staff should prepare a list of questions as a standard for all the reference sites. Once all calls are completed, meet to discuss each reference and don't move forward unless the team unanimously supports a decision to do so.
Assign a project manager for each startup. Assessing a startup can become an extended project involving many people around the organization. If these projects aren't well organized, they can drain productivity and strain customer/vendor relations. Smart companies will avoid these pitfalls by coordinating all communications and activities through a central project manager.
Visit the startup's corporate headquarters. Remember, this is not a trip to play golf with startup management and VCs. When you visit the startup's office, avoid being trapped in a cushy conference room all day. You will probably be shackled to a marketing weenie, so you need to be creative. Stop random people along the way and strike up conversations. Ask tough questions about how they help to satisfy customers. Ask about their metrics for success. While you are touring, look at the facilities for signs of disarray. Are there confidential files next to printers? Is there a lot of old-looking equipment around? These are certainly subjective assessments, but if your gut tells you that things look sloppy, catch the next plane out of town.
If your startup vendor makes it through the qualification process, it's time to put its products through the wringer. Set up a test lab that emulates your production systems. The startup should be willing to pitch in with people and time. Of course, you should test the product to its limit, but also consider:
- Usability. Once you get beyond basic functionality, is the product still intuitive or does it start to fall into a proprietary hole?
- Support. Pick up the phone and call the startup at 2 a.m. How long does it take to get a live person on the phone and to get an answer to your question? Customer support is generally rated as one of the highest priorities in IT, yet is often ignored during the product evaluation. Savvy CIOs will test and re-test startups on these capabilities before making a long-term commitment.
- Security. In storage, security is often omitted from product design in favor of performance and availability. Don't compromise here. Regulatory issues will demand strong security in the next 12 to 18 months. If startups can't adhere to your security demands, show them the door.
You can also get the startup to commit to local spares. You don't want to find out that the closest spare part is 1,500 miles away when your systems are down. If you're located outside a major metropolitan area with no local parts depot, be sure to demand that spare parts (and skilled technicians) are close by.
You should also think about escrowing all code, even though some vendors balk at this request. Make this a requirement with all software startups.
Lastly, give yourself an out. Make sure that anything the startup touches is covered by SLAs with stiff penalties. Think three strikes and you're out. This may seem harsh, but it could save you from years of living with a startup that delivers great products but poor service and support.
Startups are neither panacea nor pariah, but merely an innovative and somewhat risky alternative to traditional vendors. IT users can increase potential benefits and minimize risk by employing formal processes to assess, test, measure and do business with startups. In this case, a little bit of preparation can go a long way.
This was first published in February 2004