Shai Scharf, CEO & Co-Founder, Onaro

"If you don't know what you have, you can't predict the impact of change."

February 2, 2005

8 Min Read
Network Computing logo

Dedicated to the proposition that change is the root of all evil in a SAN, startup Onaro Inc. created a well-paying niche for itself last year.

We realized the biggest pain point had to do with change,” Onaro CEO Shai Scharf says of the impetus for Onaro's initial product. “Every time somebody changes something, there are problems. Storage area networks are so complex, they can’t deal with somebody changing what they have on them.”

So Scharf and three other Onaro founders – Roy Alon, Assaf Levy, and Raphael Yahalom, all like Scharf veterans of Israeli military intelligence who migrated to Boston (see Onaro Lifts Its Cover) – came up with SANscreen. The software is based on what they call predictive change management. SANscreen discovers devices on Fibre Channel and iSCSI networks, determines their relationships to other devices, and predicts the effect of changes on a SAN.

Onaro’s focus on change helped distinguish it from other storage resource management software vendors after SANscreen rolled out last June (see Onaro Ships Change Manager). Since then, things have moved right along. The startup lined up development partnerships with Brocade Communications Systems Inc. (Nasdaq: BRCD), Cisco Systems Inc. (Nasdaq: CSCO), Computer Network Technology Corp. (CNT) (Nasdaq: CMNT), EMC Corp. (NYSE: EMC), Emulex Corp. (NYSE: ELX), Hewlett-Packard Co. (NYSE: HPQ), Hitachi Data Systems (HDS), IBM Corp. (NYSE: IBM), McData Corp. (Nasdaq: MCDTA), and QLogic Corp. (Nasdaq: QLGC). (See Onaro Names Development Partners.) Customers have also signed on, paying $100 or more per port for SANscreen.

In 2005, Scharf says Onaro will widen its product line – and competition – by offering similar software for the entire data center (see Onaro Eyes Data Centers). He also expects to add storage management features to SANscreen. Another goal for this year is to forge new business deals. "The focus of our company in 2005 is to partner more and diversify our revenue stream," Scharf says. "There’s a lot in the works right now. We’ve achieved development partnerships where we interface with devices of our partners. Now we want to move from development relationships to revenue-generating relationships, where both parties can benefit."

Given Scharf's military intelligence background, maybe it's not surprising that when we caught up with him recently, we covered topics such as paranoia, pain, and fear. But Scharf certainly doesn't mind talking at length about Onaro's product, and grudgingly about business matters. The only subject off-limits was Cisco (see Cisco Eyes SAN Startups).

So check it out. Click on the links below for specific portions of this Byte and Switch interview:

— Dave Raffo, Senior Editor, Byte and Switch

Byte and Switch: How did Onaro get started? Scharf: We started as four co-founders, actually at the end of 2000. It wasn’t even a company yet, but we were funded by VCs to go look for a problem. It’s not like we had an idea and then went looking for money. Rather, we were financed to go look for a problem.

We have expertise in the intelligence business. The way we look for a problem is, we managed to penetrate over 100 companies and interview over 100 storage managers. We correlated the answers and pain points and combined them with our experiences in technology and intelligence.

Byte and Switch: So what did you find?

Scharf: We realized the biggest pain point had to do with change. Every time somebody changes something, there are problems. Storage area networks are so complex, they can’t deal with somebody changing what they have on them. There was a fear factor that resulted in a reluctance among storage managers to change their SANs, so they started building SAN islands instead.

We were not storage guys before, so we didn’t only look at storage. We also looked at networking, protocols, security, and other issues. We managed to perceive what is the highest pain point, and what to do with it. Then we had to figure out how difficult it is to solve, and how much money people would be willing to pay to solve it. That’s when we started hiring. Byte and Switch: So explain what predictive change management is?

Scharf: It’s the ability to understand the physical and logical characteristics of your environment. The initial area we focused on was storage area networks – we can touch on why that is later. We didn’t just focus on devices and who’s connected to who, but we wanted to understand the logical network. That is the key to connectivity. Once you understand that, you’ll be able to apply policies you have in force.

We find a discrepancy between what people think they have and what they really have. If you don’t know what you have, you can’t predict the impact. You have to understand what you have, simulate what you are about to do, and monitor the execution of what it is you have done. It’s a complete cycle. It begins with continuous troubleshooting.

Byte and Switch: Just about everybody familiar with your product agrees it is unique for storage. Do you expect others to follow?

Scharf: Only the paranoid survive – an Intel guy said that, so I have to credit him. [Ed. note: Only the Paranoid Survive is the title of former Intel CEO Andy Grove’s autobiography.] You need to decide what to do with the paranoia. Do you want to come up with new products, or are you just a one-timer? We have many innovations and have applied for patents that relate to the work we’ve performed. We use everything we learn and develop resources and intellectual property to build more things. So if somebody else catches up to where we are now, we are already somewhere else.

Our sights are set beyond storage area networks. Our next product is still in storage, but it's an enterprise solution that follows the same trend. You must be aware, be concerned, and always be alert.

Byte and Switch: You and other Onaro founders have a background in military intelligence. How did that lead you to the SAN world?

Scharf: In our work in Israeli military intelligence, we developed techniques to capture and analyze disparate information to gain insight into a covert operation. Our expertise is in how to understand the problem and approach the problem, then come in with a solution. You need to be able to take lot of pieces of data, correlate them and apply policies and then later simulate them to understand impact. Find what is the root cause of a problem, then take corrective measures. You have to take a lot of pieces of data and make sure it is all accurate. That’s what our software does on storage area networks.

Byte and Switch: How does that translate to SANscreen? Scharf: With SANscreen I can simulate a change I want to make, so before I touch my production environment I can see what execution of the change will do.

This is pure intelligence work, where you have the process of collecting data and understanding any implications relative to future scenarios.

SANscreen goes to a Brocade switch and asks, "Who’s connected to you?" The switch says, "There’s EMC on port 5 and here are the characteristics." Most people think that’s enough. We say, that’s not completely accurate, let’s go to EMC and say, "Who’s connected to you?" and EMC says, "Here’s how we’re connected," and there are discrepancies between what the two say.

That’s one example of having to initially cleanse the data and transform it into information. You take all data and correlate it endlessly, so you understand the current state of the network.

Byte and Switch: What does SANscreen tell customers about their SANs that they don't already know? Scharf: We find a discrepancy between what they think they have and what they really have. If you don’t know what you have, you can’t predict the impact of change. You have to understand what you have, simulate what you are about to do, and monitor what it is you have done. It’s a complete cycle. It begins with continuous troubleshooting.

There might be security back doors. We need to understand every port on every switch. Even if nothing is connected to that port, we have to understand its configuration. Let’s say a year ago, somebody used this port to point to an area on a server currently occupied by another application. Nobody remembers that, a year ago, this port had a different configuration. When you move a host server or storage connection from one switch to another, you need to know that all the zoning and LUN masking configurations have been updated before you connect the server to the new switch port. If it is not updated, your data can be erased.

You must understand all the relationships and possible relationships. You might have more than several million potential relationships. This is not a problem a human being can solve or remember. If you don’t understand all the relationships, you can screw things up.

Byte and Switch: After you understand all the relationships, what's the next step?

Scharf: You need to understand not just who’s connected to who but what the ramifications are and the impact you expect any changes to have. Then you have to simulate changes and monitor the execution. Byte and Switch: Between your work in intelligence and Onaro, you ran a company called Missing Link that sold voice recognition and computer telephony call center software. Were there any links from Missing Link to storage?

Scharf: No. There wasn’t much similarity, except it was software and a lot of intelligent processing of information. But nothing particularly connected to SANscreen.

Byte and Switch: What happened to Missing Link?

Scharf: It was sold. I had already started this venture.

Byte and Switch: So how did you get VCs to give you money for Onaro before you even knew what you wanted to sell? Scharf: Our VC Cedar Fund’s business model is to give pre-seed money (see Onara Scores Seed Financing

Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like

More Insights