As a result of the EIE'10 event, we are now making a small investment in an early stage Edinburgh company. So we are sorting through the details, finding a local law firm, dealing with the Scottish government grant support process, and so on. Objective being to balance getting the company started, spending as little of the initial investment as possible on the overhead of legal and operational process, while making a structure which will be attractive to additional funding.

There's no local equivalent to the Ycombinator Series AA equity financing documents - the law firms all promote the idea that each company has its own set of unique circumstances. The British Venture Capital Association documents are aimed at a much later stage investment than the one we are making, and are much more investor biased than the Y Combinator docs.

Coincidentally, Don Dodge (now at Google) and Brad Feld (Foundry Group) both posted advice for angel investors this week.
Updated to add - Will Herman posted a good list on 2 June 2010.

Fred Wilson and Chris Dixon made useful comments on termsheets for "quick closing" in 2009.

Updated (13 October 2010) to add a reference to the SeriesSeed documents. Brad Feld said yesterday that it is now cheaper to raise equity than debt, if you can find a lawyer who will use one of the sets of templates for terms.

References

Y Combinator example docs
BVCA example docs
Don Dodge - how to be an angel investor and make money
Brad Feld - suggestions for angels
Will Herman - on Angel Investing
Fred Wilson - first round term sheet
Chris Dixon - first round funding terms
Series Seed (Fenwick & West) documents

Cunning Systems evaluates product and service ideas in computing and communications. If you would like to discuss an idea, contact us at info@cunningsystems.com

The Engage Invest Exploit event put together by Edinburgh Informatics on Wednesday 12 May 2010 : a summary.

At the Informatics Forum building there was an enthusiastic audience to listen to pitches from six startups, with panel discussion on each from potential investors, including Anil Hansjee from Google, and Andrew Nutter from Balderton Capital.

Pitching companies

• M Power World (microbial fuel cell development)
• Factonomy (software development tools technology)
• Spinsight (applying vision analysis to sports coaching and statistics)
• Mobile Acuity (product recognition using mobile phone camera as input - no barcodes required)
• Satsis ( mobile phones location indoors with 5m accuracy)
• Inquisitive Systems (fraud detection)

Big news for another exhibitor, Vibio (secure sales portal system, like Ebay) - they signed a deal for $1m of funding with Archangels, a Scottish angel investment syndicate.

At the after lunch panel session I was one of four Ann(e)s on the Expert panel - exhibiting considerable experience in science, technology, investing and small company success. Prof Macrae moderated. Audience questions were dominated by the issues around intellectual property, use of patents as a proxy for merit by academic funding bodies, and the consequences for university spin out companies. As @ewanmcintosh tweeted "Very little in internet startups has to do with defensible IP. It's about doing something great, quickly & dominating mkt".

Exhibiting companies who found me to say 'see this' included

• Codeplay (programming tools for multicore processorts)
• Podfather (a delivery tracking system like FedEx or UPS use, delivered as a service )
• Blackford Analysis (technology for analysis of huge datasets, originally developed for astrophysics, applicable to oil&gas, 3D medical scans, and other parametric modeling problems)
• Trivault (de-duplication of encrypted data, backup, recovery)

The day concluded in the underground bar at Revolution on Chambers St, where I caught up with Tom McCallum, Scott Allison, and Maciej Zurawski (Musemantik).

References

Updated to add link to 10 minute video compiled from the day's events

Investor list
Exhibitor list
Preliminary meetings
Vibio deal
Michelle Rodger's summary
Mark Littlewood, BLN blog post

Cunning Systems evaluates product and service ideas in computing and communications. If you would like to discuss an idea, contact us at info@cunningsystems.com

Follow me on twitter at @annejohn and @vcwatch

Here's the first part of the report from the Engage Invest Exploit event in Edinburgh - the pre meeting meetings on Tuesday 11 May 2010.

The Informatics Forum (picture Andrew Mitchell)

Enjoyed meeting Ann Budge, one of my fellow panelists; she is now making angel investments, having retired in 2008 from Sopra, which bought her computer services company in 2005.

Immediately after that Alex van Someren talked to the Edinburgh Entrepreneurship Club about his initial approach as a teenager to Acorn, the formation of ANT, and the highlights of getting intial funding for his encryption company, nCipher. Several times in the evolution of his business life he'd faced the decision about whether to take what was on offer at the time, or hold out for something less certain later. Each time he's taken the immediate offer - going directly from high school to work for Acorn rather than going to college at Cambridge; accepting the $1m funding offer even though he gave up a controlling share of nCipher - and it has paid off.

Alex and I were dispatched in a taxi across the city to the dinner organized by Mark Littlewood of the Business Leaders Network, being held in the classic Georgian building in Queen St which is one of the premises of the Scotch Malt Whisky Society. Their attic dining room, up three flights of spiral stairs, has a fine view across the Forth estuary to the Fife hills - there was enough light left to admire it even on a damp grey evening. Dr David Milne, co-founder of Wolfson Microelectronics, was reviewing the investment landscape in Scotland as we arrived. Other participants were from Silicon Valley Bank, Seraphim Capital, Par Equity, Google, Tech Crunch, and the sponsors, Informatics Ventures. There was an instructional whisky tasting, fine food, and a good deal of business networking.

References
Edinburgh Entrepreneur Club
Investor list

Cunning Systems evaluates product and service ideas in computing and communications. If you would like to discuss an idea, contact us at info@cunningsystems.com Follow me on twitter at @annejohn and @vcwatch

On recent acquisitions by IBM, HP and Microsoft - the last note was about SAP, Google and Cisco.

IBM
Last week IBM bought Cast Iron " The combination of IBM and Cast Iron Systems will make it easy for clients to integrate business applications, no matter where those applications reside" IBM WebSphere general manager Craig Hayman said.
IBM has a substantial business integrating applications inside a business; Cast Iron has a track record of integrating applications when some part of the application support runs outside the business firewall (ie, in the cloud). "These results can be achieved using a physical appliance, a virtual appliance or a cloud service with attractive entry and pay-as-you-go pricing. "
The previous acquisition, Intelliden, (Feb 2010) offers cloud based network compliance audit. IBM would seem convinced that having solutions which operate in a location independent manner is important to its customers and its business.

HP
HP bought Palm a couple of weeks ago, giving it a mobile operating system which can be used to give the user a similar experience across multiple mobile devices. However, HP Labs, which is a reasonably reliable guide to the longer term strategic direction for HP, is, like IBM, paying a lot of attention to cloud infrastructure. It's been building an open distributed cloud computing test bed, testing federation, for new application support models. HP is hedging its bets about the importance of remote infrastructure.

Microsoft
In contrast, Microsoft's last acquisition (2 Feb 2010) was of Sentilion, a healthcare software provider. The press release mentions context management and single sign on, but says nothing about cloud infrastructure support. Oracle's latest acquisition was also in the health care space - a company called Phase Forward, which makes data management solutions for clinical trials and drug safety. Microsoft is either not getting interest in cloud solutions from its customers, or already has all the infrastructure in place which it needs.

References

IBM Cast Iron press release
HP acquires Palm press release
OpenCirrus paper
Microsoft acquires Sentillion press release
Oracle acquires Phase Forward press release

Cunning Systems evaluates product and service ideas in computing and communications. If you would like to discuss whether an idea is likely to be of interest to a large corporation, contact us at info@cunningsystems.com Follow me on twitter at @annejohn and @vcwatch

Startups need exits, otherwise known as the explanation to investors about how they are going to get a return on their investment. If there isn't a plan for an exit, then the matter under discussion isn't a business, but a hobby.

The IPO market is opening, but the majority of exits are made as a result of an acquisition. Startups can acquire other startups, but this note and at least one more is about why and how acquisitions are made by large corporations. Long established VC firms know this, have ongoing relationships with their counterparts at the corporations, which are part of the value they bring to a deal - this note is for angel investors and startup founders, who want to understand how to position their company for a corporate acquisition.

Earlier this week 2 partners from Sonnenschein hosted a discussion with Mark Gorman (Cisco), Russ Hartz (SAP),and Don Hanson (Google).

Where they find deals

SAP : in the US, (although not in Europe), there's been a recent increase in suggestions from buy side investment bankers. The majority of their deals come from the product groups as a result of their annual reviews - the target may be a specific company, or a technology request. Partners who have integrated their product with SAP improve their visibility; building relationships with the product teams improves the chances of being acquired.SAP works with the middle tier, not the top tier, banking firms. Their industry executives speak at conferences, and lay out what they are looking for - usually, things which are complementary to SAP's existing business. SAP is unlikely to acquire a company which implements things which it regards as its core competence (like ERP). Right now they are looking for cloud computing infrastructure.

Google: looking for companies who have built a great technology with a great team, especially in the mobile and social spaces. Startup CEOs should contact Google engineering leaders, who talk at conferences. Google runs competitive Challenges, which are another way of getting visibility. Plink was just acquired as a result of participating in the 2009 Android Developer Challenge. Google is not usually interested in talking to bankers.

Cisco: Looking at deals which 'move the needle' (ie, have a noticeable effect on Cisco's revenue, market share and/or annual profit numbers). Hence the Tandberg deal which just closed. Historically, Cisco's ideal target has been a company with around 100 employees, sales for revenue, around which a model can be constructed showing how the technology, if sold by Cisco, will return some multiplier to the cost to Cisco of the acquisition (deals like these can be very much smaller). The Corporate Development group is a team of people who have been investment bankers who survey the market for possible targets. They also work with the business units who are aware of gaps in Cisco's product and service offerings, and of companies which might fill those gaps. Often the relationship starts with Cisco licensing technology. Cisco also makes investments to get visibility into potential targets, putting an observer on the board. (Netsys Technologies, acquired 1996, where I was a VP, was an early example of this strategy).

References

Sonnenschein Nath, Rosenthal LLP
Cisco list of acquisitions
Google list of acquisitions
SAP list of acquisitions

Cunning Systems evaluates product and service ideas in computing and communications. If you would like to discuss whether an idea is likely to be of interest to a large corporation, contact us at info@cunningsystems.com Follow me on twitter at @annejohn and @vcwatch

There have been several announcements in the last month, which, when taken together, add up to a significant improvement in possible networked systems performance.

Starting from the inside of the server, and working outwards :

Xeon 5600
Intel announced the Xeon processor 5600 (4 and 6 core) and the i7-980 6 core processor. These support great flexibility in trading off price, frequency, power and cache size. Examples :

IBM demonstrated a 40% improvement in operations per watt in a X3650 server using these processors - see the Intel reference below for the benchmark descriptions.

2 x 10GE port Solarstorm SFN 5122F NIC card
Solarflare announced a dual port 10GE NIC card, with hypervisor bypass; Single Root I/O Virtualization (support for thousands of vNICs); stateless offload for iSCSI, RSS, TCP/IP and UDP, LSO/TSO and LRO; 5 - 8 Watts per port; driver support for Microsoft Windows®, Red Hat® and RHEL™ Linux, Solaris™, VMware®, Citrix® XenServer™ and Novell SLES. If the Enterprise Onload application acceleration is used with these NICs, " the SFN 5122F delivers sub-5 microsecond UDP/TCP application latencies, while supporting message rates in the millions. It also significantly reduces latency jitter, making it ideal for high-frequency trading." Price $1,110, due to ship in April 2010. (The press release doesn't say whether this is a one off or quantity price).

384 x 10GE ports switch
Arista announced general availability of the 7500 switch which was shown at the November SC09 supercomputing show in Portland. This is the switch architecture with buffers, so the port to port latency is given as 4.5 microseconds, with power at less than 10W per port. (The lowest latency Arista switch is the 24 x 10GE port 7124S, at 600 nanoseconds and about $500 per port). Pricing for the 7500: starts at $140,000, fully configured at $1,200 per port (which multiplies out to $460,800)

32 x 10GE load balancing, filtering, and measurement
cPacket Networks announced a device, the cVu320G, combining complete packet inspection with Marvell's Prestera™-CX Switch chips. The device will support measurement of the time taken by specific packets to move from one part of the network to another, as well as the time and contents when 10GE links are fully loaded; this kind of information about packet volume and content on the wire can be used to identify bottlenecks and inform decisions about which parts of the systems infrastructure would most benefit from upgrades. The cVu320G is due to ship in April 2010. The press release gave no indication about pricing.

Combining the latest processors with recent dual port 10GE NIC cards and high capacity, compact, low power switches provides the next generation hardware platform, for use in house, or installation in co-location facilities. Fractional shares of these platforms will soon be available from hosting providers.

The securities industry, always searching for lower latency and well as lower power, is an early adopter customer for these capabilities. NYSE Euronext announced early customer testing for "trading in a box", combining market data engines, ordering engines, smart order-routing engines and market access engines inside a single server, at the Mahwah, N.J., data center.

References
Intel Xeon 5600 press release
Solarflare NICs press release
Arista Networks press release
cPacket Networks product brief pdf
Trading in a box article

Having been asked about this more than once recently, by companies who don't yet have any users or customers :

Early stage patent applications
If you are contemplating applying for a patent for the first time for your startup company, there are several considerations.

• Getting a patent awarded typically takes several years of elapsed time.
• The obvious costs are the fees to be paid to the patent office, and the fees for a patent attorney to help with preparation.
  Search for 'cost of applying for US patent' to get approximate numbers and guidelines.
• Don't ignore the cost of your time; explaining engineering ideas to attorneys is not easy. Seriously consider whether you'd be adding more value by explaining your ideas to potential customers instead.

Brad Burnam, from Union Square Ventures, has a useful essay; in the comments, he says "In the world of web services, where we invest, you will have succeeded or failed long before a patent is approved, so it often does not make sense to invest in the process."

In addition, the question to ask is not so much what the upfront costs are, but if you could actually afford to defend a patent if you had one. Can you grow the company enough to afford settle a patent suit ?

Cisco Systems had only one patent for several years after the IPO
Cisco's approach to patents was not to spend any time or money with patent lawyers until they were about to do the IPO (in 1990). As part of that process they patented the IGRP routing protocol - and there were no more patent applications for several years, until they got to the point where they needed patents as a currency to exchange with competitors in cross-license agreements. Sequoia Capital did not require them to have applied for a patent as a pre condition of their funding. The thinking was that by the time a competitor had read the patent application and or reverse engineered the hardware, that Cisco would have developed something new and the patented product would have been superseded. Although Cisco is nominally a hardware company, in fact the hardware was perceived internally as a method for packaging software - that was the real added value.

Recently, there's been an upswing in informed opposition to software patents. Brad Feld, at Foundry, is keenly interested in the problems with the patent system and has a series of detailed posts covering the issues.

References
Brad Burnam, USV on software patents
Brad Feld, Foundry Group - series of patent posts
US patent office process chart

We went over to Stanford to see the Cool Product Expo 2010 earlier this week.

Mainly, but not exclusively, student projects, with varying degrees of sophistication in the presentation.

The people with whom we spent the most time were Sifteo - they make little blocks with a display, motion sensor, and wireless communication. The blocks are aware of each other and interact. This is a spin out from the MIT Media Lab.
They are sorting out their initial production run, and expect to have product during 2010. We discussed some of the methods for protecting intellectual property while using contract manufacturing.

Driptech have a simple gravity driven above ground irrigation system - they have targeted it at developing countries, but it seems it'd be attractive to a lot of back yard vegetable growers in this country too.

Eliptigo looked like eccentric fun - imagine a cross between an elliptical exercise machine and a bicycle, or cross country skiing with inline wheels.

Carbon fiber was a common theme - Blackbird guitars make their instruments from it; the Stanford Solar Car body is made from it.

A useful couple of hours to get perspective on new developments.

Exhibitor list

Cunning Systems evaluates product and service ideas in computing and communications. If you would like to discuss an idea, contact us at info@cunningsystems.com Follow me on twitter at @annejohn and @vcwatch

On reading Clay Shirky's complex business models collapse essay

Clay Shirky describes the process where 'old complexity' is super ceded by 'new simplicity', and how the businesses embedded in the old model have such difficulty adapting to the new cultural imperative. We all know of examples of this - however, the interesting question is the next stage. There are indications that some businesses are prepared to pay something for a differentiating amount of availability and reliability. Looking at Amazon's Web services would indicate that Amazon think so - people are wiling to pay extra, in both money and complexity, for load balancing across EC2 instances.

The customer here isn't the end user, it's the people providing the content to the end user.

Shirky essay
AWS load balancing

The Internet Operating System
Tim O'Reilly's article on the State of the Internet Operating System gathered a lot of attention. The 'operating system' metaphor is inaccurate - long ago, Cisco Marketing decided to call the software that runs on their routers IOS; that was overegged too, since there were a lot of other components which had to exist to make a functioning Internet - the article is a list of useful information gathering functions which can be run on the Internet, with some high level hints at how they might be correlated.

"Increasingly, application developers don't do low-level image recognition, speech recognition, location lookup, social network management and friend connect. They place high level function calls to data-rich platforms that provide these services. "

"There is a great opportunity for developers with vision to build forward-looking platforms that aim squarely at our connected future, that provide applications running on any device with access to rich new sources of intelligence and capability."

Identity, payment systems, and activity streams over time are functions which are getting more attention; search, advertising and, more recently, location have been well exploited.

State of IOS

Cunning Systems evaluates product and service ideas in computing and communications. If you would like to discuss an idea, contact us at info@cunningsystems.com Follow me on twitter at @annejohn and @vcwatch

Having been interested in management and measurement of the Internet since I worked on an early prototype of HP Openview, KC Claffy is a good subject for my Ada Lovelace day posting.

KC makes the Internet visible - taking what is at base a vast signaling system composed of on-off states in optics, silicon and copper, and representing it in pictures which allow people to reason about the consequences of changes.

Whois map

Doing this requires access to data measuring Internet traffic - and discovering that much of the data which is required to understand the traffic trends is either not collected, or not available. The Internet is critical infrastructure, yet the people making decisions which affect it, particularly in the public policy space, don't have the necessary information with which to make reasonable decisions. KC presented on this topic to the Department of Congress and the Federal Communications Commission - and, perhaps as a consequence, there are data collection requirements for the recipients of broadband stimulus funds.

"She is the principal investigator for the Cooperative Association for Internet Data Analysis (CAIDA) which is based at the San Diego Supercomputer Center and provides tools and analyses to promote a robust, scalable global Internet infrastructure. As a research scientist at SDSC her research interests include the collection, analysis, and visualization of workload, routing, topology, performance, and economic data on the Internet. She has been at SDSC since 1991 and holds a Ph.D. in Computer Science from UC San Diego."

Updated (9 Dec 2010) link to KC's profile.

References
Profile
Traffic Management needs presentation(pdf)
CAIDA images
About Ada Lovelace