新興開源軟件公司瞄準IPO

Can big businesses be built on the idea of giving away software free of charge?

能否靠免費分發軟件來建立大企業？

That paradox has stumped some of Silicon Valley’s smartest investors since “open source” software broke on to the tech scene, with the rise of the Linux operating system 15 years ago.

自15年前Linux操作系統崛起、“開放源代碼”（open source，簡稱：開源）軟件打破科技行業格局以來，這種看似自相矛盾的想法已經難倒了硅谷(Silicon Valley)的一些最聰明的投資者。

Red Hat, which grew by selling maintenance and support to companies that used Linux, is now worth nearly $14bn. But other hopefuls fell by the wayside — even as open-source came to infiltrate many parts of the IT world, thanks to projects such as Linux and the Apache software used in most web servers.

靠向使用Linux的企業出售維護和支持服務而發家的紅帽公司(Red Hat)，如今的市值已接近140億美元。但其他懷揣這種希望的公司卻折戟沉沙，儘管開源軟件已滲透到IT界的多個方面（這種滲透得益於在網頁服務器上廣泛應用的Linux和Apache等項目）。

Now, investors are lining up to have another go, as a new breed of open source companies report rising revenues and plan initial public offerings. And this time, they think they have better ways of turning the popularity of open source into profits.

如今，隨着新一類開源軟件公司紛紛報出不斷增長的收入並籌劃首次公開發行(IPO)，投資者正排着隊準備再次碰碰運氣。這一次，他們自認爲有更好的辦法將開源軟件的人氣轉化爲利潤。

Open source software companies rely on networks of volunteer developers to help write their code and build an initial user base. They then aim to turn users into customers, giving the software away free while selling support services or add-ons.

開源軟件公司依賴志願開發者圈子來幫助它們編寫代碼和建立最初的用戶羣。它們接下來的目標是把用戶轉化爲客戶，在免費分發軟件的同時出售支持服務或其他附加服務。

Unfortunately for the upstarts, many big tech companies have co-opted open source themselves, to fill gaps in their own technology — and have added support for the code to their own business plans.

對初創公司來說遺憾的是，很多大型科技公司已自己新上馬了開源項目，來填補自身技術上的空白，並且已把開源代碼支持服務加入到它們自己的商業計劃中。

Even Oracle and Microsoft, the most directly threatened by free software, have adjusted their strategies to accommodate aspects of open source.

即便是受自由軟件直接威脅最大的甲骨文(Oracle)和微軟(Microsoft)，也已調整自身戰略，以適應開源技術的各個方面。

However, while the first wave of open source relied on making free versions of existing programs without adding much new, Mike Volpi, a partner at venture capital firm Index Ventures says new approaches are emerging with the growth of cloud computing and “big data” analysis. He says that some developments in open source software are not being matched in the traditional world of commercial code.

不過，第一波開源浪潮依賴於生成既有程序的免費版本而不增添太多新東西，而據風投公司Index Ventures的合夥人米凱•沃爾皮(Mike Volpi)表示，隨着雲計算和“大數據”分析的發展，新的做法正不斷涌現出來。他說，開源軟件領域的一些新生事物在傳統的收費軟件界找不到對應體。

For example, Hadoop — code developed at Yahoo before it was released in open source form — is becoming a foundational technology for the era of big data. Hadoop was designed to handle the masses of data stored in the sort of distributed computing systems pioneered by large internet companies and now found in wider corporate use.

例如，Hadoop（其代碼是在雅虎(Yahoo!)開發的，而後在開源論壇發佈）正逐漸成爲大數據時代的基礎性技術之一。Hadoop旨在處理儲存在分佈式計算系統上的海量數據，這種系統的首批使用者是大型互聯網公司，如今已在企業界得到更廣泛的應用。

Earlier this year, Cloudera, one of the companies trying to build a business on Hadoop, declared itself the second open source company to hit $100m in annual revenue, after Red Hat.

今年早些時候，試圖在Hadoop平臺上建立業務的Cloudera，宣稱自己是繼紅帽之後第二家年收入達1億美元的開源軟件公司。

Open source is also proliferating in databases designed to support applications that run in the cloud. MongoDB, the standard-bearer for this wave of so-called NoSQL database companies, was valued at nearly $2bn in a round of fundraising earlier this year. Along with Cloudera — which has raised $1.2bn, much of it from Intel, and been valued privately at more than $4bn — it is at the forefront of a wave of IPOs anticipated for next year.

一些數據庫也大量採用開源技術，這些數據庫旨在支撐在雲端運行的應用。作爲這波所謂NoSQL數據庫公司中的領袖，MongoDB在今年早些時候的一輪融資中被估值爲將近20億美元。該公司與Cloudera（Cloudera已融資12億美元，其中許多來自英特爾(Intel)，私下的估值已逾40億美元）一道，均處在預期明年將到來的一波IPO潮的最前沿。

Critics, though, point to the past failure of open source developers to make much of a dent in the sector. Their technology “is not tested, it’s not hardened, and they have this new business model that’s totally unproven”, says Gary Bloom, chief executive of MarkLogic, a NoSQL company that still takes the traditional approach of selling licences to its products.

但批評人士指出，開源開發者過去並沒能在該行業中留下多少印記。MarkLogic首席執行官加里•布盧姆(Gary Bloom)稱，這些開發者的技術“未經檢驗，並不可靠，而且他們的這種新商業模式完全未經驗證”。MarkLogic是一家仍走傳統路線的NoSQL公司，依靠的是出售自己產品的許可證。

MarkLogic’s own revenues last year were “well north of $100m” and the 14-year-old company is also eying an IPO, he adds.

布盧姆補充稱，去年MarkLogic的年收入“遠超過1億美元”，而且這家創辦了14年的公司也在考慮舉行IPO。

In recent years, though, being able to reach a mass user base with free software proved a bigger lure — even if turning reach into revenue has been a challenge. “If the cost of entry to our technology was $100,000 a server, not many people would be in a position to use it,” argues Kelly Stirman, head of strategy at Mongo. There are “hundreds of thousands of deployments” of the company’s free software, he says, though it has only 2,000 paying customers.

但是，最近幾年的事實證明，能用自由軟件來與海量用戶羣建立聯繫是一個更大的誘惑——儘管如何把這種聯繫轉化爲收入一直是個難題。Mongo戰略主管凱利•斯蒂爾曼(Kelly Stirman)稱：“如果使用我們技術的成本是每臺服務器10萬美元，有財力用的人恐怕不會太多。”他說，該公司的自由軟件有“數十萬使用者”，儘管付費客戶只有2000個。

Ben Golub, chief executive of open source company Docker, suggests the new wave of developers have learnt the lessons of the past and are adding commercial tools and services to the mix.

開源軟件公司Docker的首席執行官本•戈盧布(Ben Golub)認爲，新一波的開發者已從歷史中汲取了教訓，他們正將收費工具和服務打包進來。

One opportunity lies in the need for add-on technologies to monitor and control IT systems. Mongo, for example, lets users run management software on eight servers free of charge, but makes them pay $50 a server to add more.

一個機遇在於客戶對監控IT系統的附加技術的需求。例如，Mongo允許用戶在8臺服務器上免費運行管理軟件，但用戶每增加一臺服務器需支付給它50美元。

A further opportunity comes from running the software on their own servers and providing a service to users through the cloud. Docker, for instance, sells a hosted service to small businesses that do not want to take on the work of managing their own IT.

另一個機遇在於，在它們自己的服務器上運行軟件，並通過雲向用戶提供服務。舉個例子，Docker向不願承擔管理自身IT工作的小企業提供收費託管服務。

Even so, some are still relying on the old model of offering free software but selling the maintenance and support. This is clearest in the Hadoop market, where Hortonworks — which last year became the first open source company since Red Hat to IPO — has stuck to the older business model in contrast to rivals Cloudera and MapR.

儘管如此，仍有一些公司依賴於提供自由軟件但出售維護和支持服務的舊模式。這種情況在Hadoop市場最爲明顯：與競爭對手Cloudera和MapR相比，Hortonworks仍恪守比較老的商業模式。Hortonworks去年舉行了IPO，成爲紅帽之後首家上市的開源軟件公司。

Whatever their business model, the new generation of open source companies will have to move fast to avoid being crowded out by established tech giants.

無論它們的商業模式是怎樣，新一代開源軟件公司必須迅速採取行動，以避免被老牌的科技業巨頭擠出局。

Big players’ influence is clearest in the development of OpenStack, a package of software designed as a foundation for cloud computing. IBM and Hewlett-Packard have turned to OpenStack to support their own ambitions against cloud companies including Amazon and Microsoft.

在OpenStack的開發方面，大玩家的影響力最爲明顯。OpenStack是一攬子軟件，目的是成爲雲計算的基礎。IBM和惠普(HP)已轉而使用OpenStack來支撐其對抗亞馬遜(Amazon)和微軟等雲企業的抱負。

“Everyone has an axe to grind on it,” says Mr Volpi, which has slowed progress. “There is a reasonable amount of frustration over the speed with which it is being implemented.”

“在這件事上，每個人都有自己的盤算，”沃爾皮表示，這拖慢了開源軟件的進展。“人們對開源落實的速度有些沮喪。”

It may sound like an opportunity for focused open source companies. But history suggests that even if their software catches on, commercial success may be elusive.

這聽起來也許是專注於開源軟件的公司面臨的一個機遇。但歷史告訴我們，即使這些公司的軟件受到人們的追捧，它們可能也難以取得商業上的成功。