Akamai's Linode buy: Good for enterprise, risky for others
Feisty indies that do too well don't stay indie for long
Opinion The one thing AWS can't offer is not being AWS. Google and Microsoft can, but then you're stuck with Google and Microsoft. Each of these cloud infrastructure options – AWS, GCP and Azure – are big, centralised components of bigger organisations with other things on their minds, and a deep aversion to sharing customers.
Akamai has the edge here. It has always just sold cloud infrastructure services to customers it knows shop elsewhere. It doesn't care who's hosting its fleet of enterprise clients. The stuff Akamai sells is designed to bolt onto AWS, GCP, Azure, you name it, and do the delivery to the world. The company is intrinsically multi-platform. With last week's Linode acquisition, Akamai has as little problem adding hosting, storage, and developer-friendly cloud services to its portfolio as a supermarket does adding own-brand cornflakes to its shelves.
Akamai is no supermarket. It's a highly competent, highly trusted deliverer of high-availability network services to rich companies, and it has the sales and support network to match. Linode's chief attraction to Akamai wasn't primarily its technologies or existing infrastructure, things that Akamai could do for itself out of parts it already has at home. Rather, it was Linode's reputation among its customers, which closely matches what Akamai has among its client base. That it's up and running is great and gives Akamai a near-instant entry to the cloud compute provider's club, but it's Linode's decades of trust with developers that's the golden ticket.
The best we can hope for is that it sees the customer-facing side of the existing Linode organisation as it does the other partners through which it sells developer-focused services...
Akamai – now with added Linode! – has lots of options. It's simultaneously the plucky underdog with revenues a 30th of AWS's and the bluest of blue-chip enterprise-scale service providers. It can take Linode and package it as the Multiplatform-O-Tron that enterprises need to reduce reliance on single suppliers. Better deals, more uptime, same great taste. It can offer complete core-to-edge single-billed infrastructure, with complete developer tools and support to match.
In so doing, of course, it moves compute-at-scale much closer to the edge, which is a highly fashionable option right now, as Amazon well knows.
Building a highly reliable compute and storage infrastructure is hard. Building a global high-performance redundant network is hard. You need both to compete with the big guys. Akamai's there, and it's going to be there as an incumbent wherever Amazon tries to get closer to the edge.
This is all very exciting – except, perhaps, for some of the existing Linode customers. Linode customers are diverse, technical, and very loyal. It's typically the platform of choice for startups with a good idea, Linux chops, and a dislike of the herd. It's the cloud provider that delivers superb bang-per-buck to those who know what they're doing and don't want to do anything else. Linode users do innovative things, they're savvy, they're everywhere, and – currently – you could argue that they're completely invisible to Akamai. Linode has free networking tools available via nmap.org, Akamai has adobe.com. It's a huge cultural divide.
- The end of free Google storage for education
- Google Cloud takes a gap year. It may come back with very different ideas
- Machine learning the hard way: IBM Watson's fatal misdiagnosis
- IPv6 is built to be better, but that's not the route to success
It's not that Linode doesn't have a presence in some big companies; you can go a long way as a small team building your particular service that way. But you won't be part of the big picture, you'll be proof of concept. Akamai has heard of small and medium-sized companies; it encourages ISPs to sell security products to them. But until the deal goes through, Akamai isn't the sort of company with a $5/month entry tier, and what's the chance it'll want to be one afterwards? It'll have tiers, but they'll all come with Akamai services and you can't support those properly on single-digit plans.
It would be a great mistake if Linode becomes purely Akamai's cloud service platform for its enterprise customers. Linode is a great business that's grown strongly into a multi-continent near-billion-dollar cloud provider with lots of customers who love it. It has done that in the face of trillion-dollar competition by filling a real need, a need that may now go unmet.
It's the role Linode plays in the ecosystem that's so important. Like a successful indie record label, it identifies with its customers and talent. If you don't want to fund Amazon or Google or Microsoft because you care about such things, Linode is there for you.
The tech world, and the cloud infrastructure that increasingly defines IT, show all the evidence of monocultural susceptibility to systemic malaise. Google has given up not being evil. Amazon has the appetite of a monster. Working at scale has a dreadful lack of options, of diversity, and we desperately need more of both. Linode's capture in the name of enterprise will not help.
It doesn't look good. Akamai talks of all the benefits it sees to its product and service line-up, its potential for growth, its improved competitive stance, and the importance to its enterprise customers. It hasn't mentioned the relevance to nmap.org.
The best we can hope for is that it sees the customer-facing side of the existing Linode organisation as it does the other partners through which it sells developer-focused services, its own Digital Ocean. That would fulfil most Linode customers' needs, to be left alone to get on with things – ah, what a butterfly dream it is, and how often enterprise-level thinking crushes it on the wheel.
It is good that there is the potential for better engineered robust cloud services at scale, adding multi-platform, decentralised, secure components that remove single points of failure. If that comes at the cost of the creativity and independence that made it possible in the first place, that cost will be too high. ®