By William Fellows
Trying hard not to spill the beans on a new release of its internet caching system (ICS) but unable to contain its enthusiasm, Novell Inc says a 2.0 release of the code due by mid- next year will be the building block for new volume applications that should present it with a significant new revenue opportunity. Initially Novell will integrate programs for managing and monitoring performance and bandwidth and for creating personal identity servers that could cache employee or customer profiles like those created with its DigitalMe system introduced last week. It is also working on content distribution services with Akamai.
Although the company will leverage its Novell Directory Services on top of ICS wherever practicable, NDS is not a prerequisite for using ICS, which Novell sees as a very valuable volume opportunity. Nevertheless the content distribution technology it is working on with Akamai will be offered in three pieces: and why not manage in a directory like NDS! Novell says.
It already has the likes of Dell, IBM, NEC and Compaq bundling ICS on their servers and sees ICS as largely an OEM play. It hopes they will pick up its forthcoming applications as well as those developed for the software by partners including Edgix and N2H2.
Novell says ICS 1.1, which is due imminently, offers better latency, bandwidth and transmission costs than all other caching systems, including those from chief rivals Inktomi and Network Appliance. It admits Inktomi is more geared towards back-end integration support, but believes it is still very expensive compared with ICS. Novell claims the Cache Object Store at the heart of ICS enables it to operate at ten times the speed of rival systems. For any company delivering a T1 pipe or more of HTML traffic Novell claims ICS offers at least a 30% saving on bandwidth, not to mention the associated savings on employee time.
But aren’t caching systems going to become redundant as broadband net access takes off? No, says Novell, which believes that every available bit of bandwidth is going to be consumed as soon as it becomes available, which will make caching systems ubiquitous.
To our minds the proliferation of caching systems is a necessary evil at least until broadband access morphs them into more value- added application services. A company buys a web server, sets up intranet, extranet, e-commerce and firewall applications – or outsources it. Then it finds it needs to buy or rent another server to store regularly accessed information just so that it can get decent network speed. It’s a great way to sell the customer another box – and market projections says that 80% of caching revenue is in hardware sales – but that they are needed at all just seems plain wrong. The market may be converging around infrastructure and diverging on appliances, but caching currently falls somewhere in between. The Internet Research Group estimates that the internet caching market will be worth $2.2bn by 2003.
Moreover, now that Novell is licensing NDS fast and furiously to sew up the market before Microsoft Active Directory arrives, and to provide it with a large installed base on top of which it can sell directory-enabled servers, why doesn’t it turn NDS over to an open source license and development model? Such a move is still under consideration even though its chief proponent SVP Chris Stone has left the company. Novell says its main reservation is the fear that open sourcing NDS would result in a rash of incompatible versions of the software.
